WO2021046775A1

WO2021046775A1 - Resource scheduling, application, and control methods, apparatus, system, and storage medium

Info

Publication number: WO2021046775A1
Application number: PCT/CN2019/105456
Authority: WO
Inventors: 陶云峰; 卢毅军; 宋军; 唐治洋; 奉有泉; 冯亦挥
Original assignee: 阿里巴巴集团控股有限公司
Priority date: 2019-09-11
Filing date: 2019-09-11
Publication date: 2021-03-18
Also published as: CN114008996A

Abstract

Embodiments of the present application provide resource scheduling, application, and control methods, an apparatus, a system, and a storage medium. In the embodiments of the present application, an intermediate layer is added in between an application layer and an IT resource layer, and a resource bidding and control mechanism is used in the intermediate layer, so as to combine resource requirements of the application layer to perform resource scheduling with respect to the IT resource layer. The invention thus enables reasonable resource allocation, improves the overall resource utilization rate, and promotes overall energy efficiency.

Description

Resource scheduling, application and control method, equipment, system and storage medium

Technical field

This application relates to the technical field of data centers, and in particular to a resource scheduling and application method, device, system, and storage medium.

Background technique

With the development of cloud computing technology, various computer room systems or data center (Data Center, DC) systems are constantly being deployed, and the pressure carried by the computer room system or data center system continues to increase. Can the computer room system or data center system be used rationally? IT equipment, improving the utilization rate of IT equipment, directly affects the overall energy efficiency of the computer room system or data center system.

In the prior art, IT equipment is generally scheduled according to the priority of the application, the load of the IT equipment, and other factors. For example, IT equipment is assigned to higher priority applications first, or the application is assigned to higher load applications. Light IT equipment. This resource scheduling method is simple and easy to implement, but the overall resource utilization rate of the computer room system or data center system is not high, and the overall energy efficiency is not high.

Summary of the invention

The embodiments of the present application provide a resource scheduling, application, and control method, device, system, and storage medium, which are used to improve overall resource utilization and improve overall energy efficiency.

The embodiment of the application provides a system, including: an application layer, an intermediate layer, and an IT resource layer; the intermediate layer is between the application layer and the IT resource layer, and is used to adopt a resource bidding and regulation mechanism to combine The resource requirements of the application layer perform resource scheduling on the IT resource layer.

An embodiment of the present application provides a resource scheduling method, including: obtaining resource requirements of upper-layer applications; using a resource bidding and regulation mechanism, combined with the resource requirements of the upper-layer applications, to perform resource scheduling on lower-layer IT resources.

The embodiment of the present application provides a resource control method, including: monitoring the status information of the schedulable resources participating in the resource bidding in the resource market; according to the status information of the schedulable resources participating in the resource bidding in the resource market, The schedulable resources participating in the resource bidding are regulated.

An embodiment of the present application also provides a resource application method, including: determining at least one task that a first application needs to complete and resource information that the at least one task needs to use; and determining the resource information that needs to be used by the at least one task The actual bid value given for the resource required for the at least one task application; generating a resource bid request based on the resource information required for the at least one task and the actual bid value given for the resource required for the at least one task application ; Send the resource bidding request to the resource scheduler for the resource scheduler to bid for the first application and other applications for schedulable resources; receive the first application returned by the resource scheduler The result of the program bidding application for schedulable resources.

The embodiment of this application also provides a resource application method, including: determining the number of resources that the first application needs to apply for this time and the total budget price of this resource application; according to the number of resources that need to be applied this time and the amount of this resource application Budget total price, generate resource bidding request; send the resource bidding request to the resource scheduler, so that the resource scheduler can bid for the first application and other applications to apply for schedulable resources; receive the resource scheduling The result of the first application bidding application for schedulable resources returned by the party.

The embodiment of the present application also provides a resource scheduling method, including: receiving a resource bid request, the resource bid request including: the number of resources that the first application needs to apply for this time and the total budget of the resource application; The number of resources that need to be applied for and the total budgeted price of this resource application. Among the schedulable resources participating in the resource bidding, determine at least one resource to be applied for and its actual bid value; according to the actual bid value of the at least one resource to be applied for, Apply for the at least one resource to be applied for by bidding for the first application program with other application programs, and return the resource application result of the first application program.

An embodiment of the present application also provides a resource control device, including: a memory and a processor; the memory is used to store a computer program; the processor is coupled with the memory and is used to execute the computer program for use In: monitoring the status information of the schedulable resources participating in the resource bidding in the resource market; according to the status information of the schedulable resources participating in the resource bidding in the resource market, regulating and controlling the schedulable resources participating in the resource bidding.

The embodiment of the present application also provides a resource application device, including: a memory, a processor, and a communication component; the memory is used to store a computer program; the processor is coupled with the memory and is used to execute the computer program , For: determining at least one task that the first application needs to complete and resource information that the at least one task needs to use; according to the resource information that the at least one task needs to use, it is determined to be required by the at least one task application The actual bid value given by the resource; generate a resource bid request based on the resource information needed to be used by the at least one task and the actual bid value given for the resource required for the at least one task application; The resource bidding request is sent to the resource dispatcher for the resource dispatcher to bid for the first application and other applications to bid for schedulable resources; and to receive the first application bidding application returned by the resource dispatcher The result of schedulable resources.

The embodiment of the present application also provides a resource application device, including: a memory, a processor, and a communication component; the memory is used to store a computer program; the processor is coupled with the memory and is used to execute the computer program , To: determine the number of resources that the first application needs to apply for this time and the total budget price of this resource application; generate resource bid requests based on the number of resources that need to be applied this time and the total budget price of this resource application; Send the resource bid request to the resource scheduler through the communication component, so that the resource scheduler can bid for the first application and other applications for schedulable resources and return; and receive the resource scheduler The returned result of the first application bidding for schedulable resources.

An embodiment of the present application also provides a resource scheduling device, including: a memory, a processor, and a communication component; the memory is used for storing a computer program; the processor is coupled with the memory and is used for executing the computer program , For: receiving a resource bid request through the communication component, the resource bid request including: the number of resources that the first application needs to apply for this time and the total budget of this resource application; the resources to be applied for according to this need The actual bid value of at least one resource to be applied for is determined by the quantity and the total budget price of this resource application; according to the actual bid value of the at least one resource to be applied for, the bid application for the first application and other applications At least one resource to be applied for, and the resource application result of the first application is returned through the communication component.

An embodiment of the application also provides a resource management system, including: a resource scheduler, a resource demander, and a resource regulator; the resource demander is used to generate a resource bid request for a first application program; and the resource bid request Sent to the resource scheduler to participate in resource bidding; the resource scheduler is used to bid for the first application and other applications among the schedulable resources participating in the resource bidding according to the resource scheduling request Apply for schedulable resources; the resource regulator is used to monitor the status information of the schedulable resources participating in the resource bidding in the resource market; according to the status information of the schedulable resources participating in the resource bidding in the resource market, The schedulable resources participating in the resource bidding are regulated.

An embodiment of the application also provides a computer room system, including: at least one schedulable resource, at least one infrastructure, resource application equipment, resource scheduling equipment, and resource control equipment; the at least one infrastructure is the at least one schedulable resource Provide basic services; the resource demand device is used to generate a resource bid request for the first application; send the resource bid request to the resource scheduling device to participate in the resource bid; the resource scheduling device is used to The resource scheduling request, among the schedulable resources participating in the resource bidding, applies for schedulable resources for the first application program and other application programs for bidding; the resource control device is used to monitor the availability of the schedulable resources participating in the resource bidding State information in the resource market; according to the state information of the schedulable resource participating in the resource bidding in the resource market, regulating and controlling the schedulable resource participating in the resource bidding.

An embodiment of the present application also provides a data center system, including: at least one computer room, resource application equipment, resource scheduling equipment, and resource control equipment; each computer room includes at least one schedulable resource and at least one infrastructure. The facility provides basic services for the at least one schedulable resource; the resource demand device is used to generate a resource bidding request for the first application program; and the resource bidding request is sent to the resource scheduling device to participate in the resource bidding; The resource scheduling device is configured to, according to the resource scheduling request, apply for schedulable resources for the first application program and other application programs among the schedulable resources participating in the resource bidding; the resource control device is configured to Monitoring the status information of the schedulable resources participating in the resource bidding in the resource market; according to the status information of the schedulable resources participating in the resource bidding in the resource market, regulating and controlling the schedulable resources participating in the resource bidding.

The embodiments of the present application also provide a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the processor is caused to implement the steps in the method embodiments of the present application.

In the embodiment of the present application, an intermediate layer is added between the application layer and the IT resource layer, and the resource bidding and regulation mechanism is adopted in the intermediate layer, and resource scheduling is performed on the IT resource layer in combination with the resource requirements of the application layer, so that resources can be performed more reasonably. Allocation is conducive to improving the overall resource utilization rate, thereby improving overall energy efficiency.

Description of the drawings

The drawings described here are used to provide a further understanding of the application and constitute a part of the application. The exemplary embodiments and descriptions of the application are used to explain the application, and do not constitute an improper limitation of the application. In the attached picture:

FIG. 1a is a schematic diagram of the structure and working principle of a system provided by an exemplary embodiment of this application;

FIG. 1b is a schematic diagram of the structure and working principle of another system provided by an exemplary embodiment of this application;

FIG. 1c is a schematic flowchart of a resource scheduling method provided by an exemplary embodiment of this application;

FIG. 1d is a schematic diagram of the structure and working principle of a resource management system provided by an exemplary embodiment of this application;

FIG. 1e is a schematic diagram of the usage state of servers on two cabinets in a computer room provided by an exemplary embodiment of this application;

2 is a schematic structural diagram of a computer room system provided by an exemplary embodiment of this application;

FIG. 3 is a schematic structural diagram of a data center system provided by an exemplary embodiment of this application;

FIG. 4a is a schematic flowchart of a resource control method provided by an exemplary embodiment of this application;

FIG. 4b is a schematic flowchart of a resource application method provided by an exemplary embodiment of this application;

FIG. 4c is a schematic flowchart of another resource application method provided by an exemplary embodiment of this application;

FIG. 4d is a schematic flowchart of a resource scheduling method provided by an exemplary embodiment of this application;

FIG. 5a is a schematic structural diagram of a resource control device provided by an exemplary embodiment of this application;

FIG. 5b is a schematic structural diagram of a resource application device provided by an exemplary embodiment of this application;

FIG. 5c is a schematic structural diagram of another resource application device provided by an exemplary embodiment of this application;

Fig. 5d is a schematic structural diagram of a resource scheduling device provided by an exemplary embodiment of this application.

detailed description

In order to make the objectives, technical solutions, and advantages of the present application clearer, the technical solutions of the present application will be described clearly and completely in conjunction with specific embodiments of the present application and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of this application.

In view of the low overall resource utilization rate of the existing computer room system or data center system, and low overall energy efficiency technical problems, in some embodiments of the present application, an intermediate layer is added between the application layer and the IT resource layer, and the intermediate layer is adopted The resource bidding and regulation mechanism, combined with the resource requirements of the application layer to perform resource scheduling on the IT resource layer, can allocate resources more reasonably, which is conducive to improving the overall resource utilization rate and thus the overall energy efficiency.

The technical solutions provided by the embodiments of the present application will be described in detail below with reference to the accompanying drawings.

FIG. 1a is a schematic diagram of the structure and working principle of a system provided by an embodiment of the application. As shown in Figure 1a, the system is logically divided into: application layer, middle layer and IT resource layer; among them, the middle layer is between the application layer and the IT resource layer.

The IT resource layer is mainly responsible for providing various IT resources, such as cabinets, servers, and CPU, GPU, memory, network cards and other resources in the servers. It should be noted that the IT resource in this embodiment may be a pure hardware resource form, or may be a resource form including hardware and software related to the normal operation of the hardware. Take the server in IT resources as an example. The server includes hardware resources such as CPU, GPU, network card, memory, and hard disk at the same time. It can also contain software resources such as drivers for these hardware resources and operating systems.

The application layer includes at least one application program, and the operation of these applications depends on the IT resources provided by the IT resource layer. For the application layer, IT resources provided by the IT resource layer can be applied for according to resource requirements. From the perspective of the application layer, more attention is paid to the performance brought by IT resources, but there is a lack of motivation to pay attention to other aspects of IT resources, such as the power consumption of IT resources.

In order to allow the application layer to simultaneously consider the performance brought about by IT resources and other aspects of IT resources when applying for resources, in this embodiment, an intermediate layer is added between the application layer and the IT resource layer, and the intermediate layer uses resources. The bidding and control mechanism, combined with the resource requirements of the application layer, performs resource scheduling on the IT resource layer. The other aspects of IT resources are related to the competitive value of IT resources, and the bidding mechanism can stimulate the application layer to pay attention to other aspects of IT resources besides performance.

Taking the power consumption of IT resources as an example, the power consumption of IT resources is related to the competitive value of IT resources. The bidding mechanism can stimulate the application layer to pay attention to the power consumption of IT resources, which can allocate IT resources more reasonably. On the other hand, it can also improve the overall efficiency of IT resources.

It is explained here that the application program in this embodiment has a broader meaning, generally refers to any program code written for the application purpose of the user, including but not limited to: system application, desktop application, driver application, network application Programs, mobile phone applications, Internet of Things applications, etc. In terms of product form, these applications can be APP, client, cloud service, application system, business system, a certain functional module or unit in business system, etc.

It should be noted that, in the embodiments of the present application, the implementation of the intermediate layer is not limited. For example, from the perspective of software and hardware, the intermediate layer of this embodiment can be implemented in a pure software form, can also be implemented in a combination of software and hardware, or even implemented in pure hardware.

In addition, whether it is a software implementation, a hardware implementation, or a combination of software and hardware implementation, the embodiment of the present application does not limit the internal structure of the intermediate layer. All internal implementation structures that can use resource bidding and scheduling mechanisms to perform resource scheduling on the IT resource layer are applicable to the embodiments of the present application. In the following embodiments, an implementation structure of the intermediate layer will be given.

Fig. 1b is a schematic diagram of the structure and working principle of another system provided by an exemplary embodiment of this application. As shown in Figure 1b, the system includes: an application layer, an intermediate layer, and an IT resource layer; wherein, the intermediate layer is between the application layer and the IT resource layer. Among them, the IT resource layer includes at least one schedulable IT resource. In Figure 1b, taking multiple servers as an example, schedulable IT resources are illustrated.

As shown in Figure 1b, the middle layer includes: a resource scheduling node and a resource control node. From the perspective of its role, the resource scheduling node is closer to the role of the market (Market), and can also be called the market node.

Among them, the resource control node is mainly used to adopt a resource control mechanism to control IT resources that can participate in resource bidding in the IT resource layer. In other words, the resource control node can determine how many IT resources are in the IT resource layer and which IT resources are eligible to enter the resource market to participate in resource bidding; it can adjust the number of IT resources that can participate in resource bidding in the IT resource layer, for example, Increase or decrease; the competitive value of IT resources that can participate in resource bidding in the resource market can also be adjusted, for example, it can be increased or decreased; and so on.

For resource scheduling nodes, a resource bidding mechanism can be used to schedule IT resources in the IT resource layer that can participate in resource bidding in combination with the resource requirements of the application layer. For IT resources that have not participated in the resource bidding, the resource scheduling node cannot schedule these IT resources.

Under the resource bidding mechanism, applications at the application layer have a resource account. The resource account stores application resources, which can be virtual funds, actual funds, credit values, points, energy values, or equity values. Among them, the resources in the resource account can be used to request access to any type of IT resources that exist in the resource market, such as CPU, hard disk, memory, network IO, and so on. Among them, these IT resources entering the resource market come from the IT resource layer.

Under the resource bidding mechanism, the resource scheduling node can schedule IT resources that can participate in the resource bidding according to the following principles. First, for any application, before using a certain IT resource, it needs to use the resources in the resource account to request the resource scheduling node to obtain its usage rights, that is, the resource scheduling node needs to request the IT resource according to the request of the application. Perform scheduling to determine whether the right to use the IT resource can be given to the application. Second, when multiple applications request the resource scheduling node to obtain the right to use the same resource, the resource scheduling node needs to allocate the resource to one of the applications according to the bidding rules. If the bidding rules are for the higher bidder, the resource scheduling node needs to allocate the right to use the resource to the application with the highest bid among multiple applications. If the bidding rule is the one with the lowest price, the resource scheduling node needs to allocate the right to use the resource to the application with the lowest bid among multiple applications. Third, under normal circumstances, when an application requests a resource that is being used by another application from the resource scheduling node with a higher competitive value, the other application should release the resource, if the other application refuses to release it , Then it is necessary to request the resource from the resource scheduling node with a higher competitive value in order to continue to occupy the resource.

In some exemplary embodiments, as shown in FIG. 1b, in addition to a resource scheduling node and a resource regulation node, the middle layer also includes a resource manager. The resource manager is used to manage the IT resources in the IT resource layer, for example, it can control the power consumption of the IT resources in the IT resource layer.

The resource manager can monitor the load of each IT resource in the IT resource layer, and control the power consumption of each IT resource according to the load of each IT resource. For example, if it is found that a certain IT resource (take a server as an example) has no load at the time, and it is predicted that the server will not have any load in the future period, the server can be shut down, or the server can be controlled to enter a certain level of low Power consumption mode, or you can reduce the CPU frequency of the server, etc.

Further alternatively, considering that the number of IT resources may be large and the load situation is more complicated, artificial intelligence and power consumption control can be combined, and the machine learning model can be used to learn the relationship between the load and power consumption of each IT resource, and then according to the learned The relationship between the load and power consumption changes, and the power consumption control of each IT resource. In order to facilitate description and distinction, the machine learning model adopted by the resource manager is referred to as the first machine learning model.

In addition to controlling the power consumption of IT resources, the resource manager can also interact with the resource control node to feed back the running status of the related IT resource to the resource control node. For example, the resource manager may feed back to the resource control node the running status of the IT resource each time the power consumption control operation is executed. For another example, the resource manager may also feed back to the resource control node the operating status of the IT resources required by the resource control node for control according to the request of the resource control node. Here, the running state of the IT resource may include the running state of the IT resource in a historical period, and may also include the running state of the IT resource in a future period. The operating status of IT resources in the future includes but is not limited to: IT resources will enter low power consumption mode for 5 minutes, IT resources will work at a lower operating frequency for 30 minutes, IT resources will be shut down for 15 minutes, and so on.

For the resource control node, the resource control mechanism can be used to control the IT resources that can participate in the resource bidding in the IT resource layer based on the operating status of the related IT resources fed back by the resource manager.

Further, considering the complexity of the regulation process, artificial intelligence technology can be combined to regulate the IT resources that can participate in resource bidding in the IT resource layer. Based on this, the resource control node can use the second machine learning model to learn the change law of IT resources participating in resource bidding in combination with the operating status of related IT resources fed back by the resource manager; use the resource control mechanism to control IT based on the learned change law IT resources in the resource layer that can participate in resource bidding. Among them, the law of change learned by the second machine learning model includes but is not limited to: the number of IT resources that need to be increased in a certain period of time to participate in resource bidding, the amount to be increased, and which IT resources can be added to the market to participate in the bidding Etc.; or, in a certain period of time, it is necessary to reduce the number of IT resources that can participate in the resource bidding, the amount to be reduced, and which IT resources can be withdrawn from the market, etc.; or, when the conditions are met, it is necessary to adjust the participation in the resource bidding The competitive value of IT resources in the resource market, how to adjust, how much to adjust, etc. Among them, the resource control node uses machine learning models in multiple places, and these machine learning models can be multiple or the same machine learning model, which is not limited.

Further, in still other exemplary embodiments, as shown in FIG. 1b, the intermediate layer of this embodiment further includes: a proxy node. The proxy node is mainly provided for applications in the application layer, and is mainly used to initiate a resource bidding request to the resource scheduling node for the application in the application layer to apply for resources. The number of proxy nodes can be one or multiple, depending on the number of applications that need to use proxy nodes. For any application, if you need to use a proxy node, you can apply to the middle layer, and the middle layer can deploy the corresponding proxy node for the application. In one implementation, the agent node may be a Broker (literally translated as a broker) node, but it is not limited to this.

It should be noted that the application program can initiate a resource bidding request to the resource scheduling node through the agent node to apply for resources. In simple terms, it is to request the resource scheduling node to purchase resources through the agent node. Of course, the application program can also directly initiate a resource bidding request to the resource scheduling node to apply for resources.

Among them, with the help of the advantages of proxy nodes, the resource information required by the application can be analyzed more comprehensively, intelligently, and reasonably, so that while ensuring system performance, it can further reduce the cost of purchasing resources. Further, if the cost is related to the power consumption of IT resources, it is beneficial to select IT resources with lower power consumption costs, thereby reducing the overall power consumption of the system.

Optionally, in order to achieve the above purpose, artificial intelligence technology can be used, using a third machine learning model to learn the operating rules of the application program corresponding to the agent node. Based on the operating rules of the application program, it can be known that the application program needs to complete at least A task and the resource information that these tasks need to use, etc.; based on the resource information that at least one task needs to use, a resource bidding request for the application can be generated. According to different resource bidding methods supported by the resource scheduling node, the information contained in the resource bidding request will also be different.

Further, the resource manager can also feed back the running status of related IT resources to the agent node to assist the agent node in generating resource bidding requests for the application. For example, the agent node can refer to the running status of related IT resources when giving competitive value for the resources required by the application. For example, if some IT resources will be closed in the future, it means that the number of such IT resources in the resource market will be reduced, and the agent node can appropriately increase the competitive value.

It should be noted that FIG. 1b shows only an exemplary implementation structure of the intermediate layer, and is not limited to this. In addition, in the embodiment shown in FIG. 1b, the working principle of each node or unit in the implementation structure is mainly introduced in general, and the specific or detailed implementation mode is not limited, and everything falls within the scope of the above general introduction. Various detailed implementation manners are applicable to the embodiments of the present application. For example, with regard to the technical means used by the resource control node to achieve the purpose of controlling IT resources to participate in resource bidding or not to participate in resource bidding, and the specific implementation methods for regulating IT resources that can participate in resource bidding in the IT resource layer, regarding the resource scheduling node Specific implementation manners for scheduling IT resources participating in resource bidding are not limited in the embodiments of this application, and exemplary solutions will be given in subsequent embodiments.

Based on the above system, a resource scheduling method suitable for the middle layer is shown in Figure 1c. The resource scheduling method includes:

11c. Obtain the resource requirements of upper-layer applications.

12c. Use resource bidding and regulation mechanisms, and combine the resource requirements of upper-layer applications to perform resource scheduling on lower-layer IT resources.

Further, an implementation manner of step 12c includes: adopting a resource regulation mechanism to regulate the IT resources that can participate in resource bidding in lower-level IT resources; adopting a resource bidding mechanism, combined with the resource requirements of upper-layer applications, can participate in resource bidding for lower-level IT resources IT resources for scheduling. It should be noted that resource bidding and resource regulation are two parallel operations in the resource scheduling process. The two operations will affect each other, and there is no strict order in execution.

Further, in the process of adopting a resource control mechanism to regulate the IT resources in the lower-level IT resources that can participate in the resource bidding, the operating status of the related IT resources in the lower-level IT resources can be combined to regulate the IT resources in the lower-level IT resources that can participate in the resource bidding. For example, if some IT resources will be closed in the future, it means that the number of such IT resources in the resource market will be reduced. When requesting such IT resources, the competitive value can be appropriately increased.

It should be noted that in the above system and method embodiments, it is possible to further combine power consumption control and resource bidding and regulation mechanisms, with the goal of reducing power consumption, using resource bidding and regulation mechanisms to schedule and regulate IT resources, With the help of reasonable scheduling and regulation of IT resources, the goal of reducing overall power consumption and improving overall energy efficiency is achieved.

In the following exemplary embodiments of the present application, a resource management system 100 will be discussed. The resource management system 100 adopts a resource bidding and regulation mechanism to perform management such as scheduling and regulation on the resources it manages. The resource management system 100 can be used as a specific implementation of the intermediate layer in the foregoing embodiment, and can use a resource bidding and control mechanism to schedule IT resources in the IT resource layer in combination with the resource requirements of the application layer. However, the resources that the resource management system 100 can manage are not limited to IT resources in the IT resource layer. In the following exemplary embodiments of the present application, resources that can be managed by the resource management system 100 are referred to as schedulable resources. The management here includes, but is not limited to: application, scheduling, regulation, and release of schedulable resources.

Among them, schedulable resources refer to resource devices, modules, or units that can provide certain services, functions, or resources (for example, computing resources, storage resources, or network resources) to the outside. From the perspective of resource form, schedulable resources can be IT resources, for example, physical devices such as servers or printers, or hardware modules such as CPU, GPU, memory, network cards, or hard disks on the server. Of course, schedulable resources can also be software resources, for example, software modules or units such as virtual machines and cloud services running on physical devices. Generally speaking, the server, as a resource form, is a relatively common schedulable resource.

In some alternative embodiments, the schedulable resources managed by the resource management system 100 are resources in the resource management system 100. In other alternative embodiments, the schedulable resources managed by the resource management system 100 are resources in other systems other than the resource management system 100; other systems are authorized to entrust the resource management system 100 to manage their resources. In still other alternative embodiments, the schedulable resources managed by the resource management system 100 may also come from multiple systems, and multiple systems have an assisting relationship and can share resources. The schedulable resource managed by the resource management system 100 may be one or multiple.

In order to better manage schedulable resources, as shown in FIG. 1d, the resource management system 100 includes: a resource scheduler 101, a resource demander 102, and a resource regulator 104. The resource scheduling party 101, the resource demanding party 102, and the resource regulating party 104 cooperate to perform resource scheduling and regulation based on the resource bidding and regulation mechanism, which can make the resource scheduling more reasonable and help improve the overall resource utilization.

Among them, the resource demander 102 represents a party with resource demand, which can be an application with resource demand, or a functional module, SDK, plug-in, or independent service responsible for making resource application for an application with resource demand. There may be one or more resource demanders 102, and they are mainly responsible for applying for schedulable resources from the resource dispatcher 101 in a bidding manner. The resource demander 102 is similar to the application program in the application layer. The application program with resource demand here also has a broader meaning, and generally refers to any program code written for the application purpose of the user.

Among them, the resource scheduling party 101 mainly adopts a resource bidding mechanism to schedule schedulable resources participating in the resource bidding. In terms of implementation form, the resource scheduling party 101 may be an application program, a business system, a service, or a hardware device with resource scheduling capabilities, which is not limited. In this embodiment, at least one schedulable resource managed by the resource management system 100 is divided into two types, one is schedulable resources that participate in resource bidding, and the other is schedulable resources that fail to participate in resource bidding. These two types of schedulable resources change dynamically, and this dynamic change is controlled by the resource regulator 104. The resource control party 104 mainly adopts a resource control mechanism to control schedulable resources participating in resource bidding.

It should be noted that the resource control party 104 regulates the schedulable resources participating in resource bidding based on the resource control mechanism, and cooperates with the resource demander 102 and the resource dispatcher 101 to implement the resource application and scheduling process based on the resource bidding mechanism. Are two parallel operations. The process of resource regulation and the process of resource application and scheduling will be respectively described below.

Resource regulation based on resource regulation mechanism:

For the resource control party 104, it can monitor the status information of the schedulable resources participating in the resource bidding in the resource market; according to the status information of the schedulable resources participating in the resource bidding in the resource market, perform the schedulable resources participating in the resource bidding Regulation. Among them, the resource regulating party 104 regulates and regulates the schedulable resources participating in the resource bidding, and can assist the resource scheduler 101 in resource scheduling, which is beneficial to the resource scheduler 101 to perform resource scheduling more reasonably.

Among them, the state information of the schedulable resource participating in the resource bidding in the resource market can represent various states of the schedulable resource in the resource market, mainly including the running state and the bidding state. Correspondingly, the status information includes, but is not limited to: resource parameters that characterize the operating status of the schedulable resource in the resource market, and bidding parameters that characterize the bidding status of the schedulable resource in the resource market.

The resource parameters include, but are not limited to: resource parameters related to the performance of schedulable resources, and resource parameters related to the power consumption of schedulable resources. Depending on the implementation form of the schedulable resource, the resource parameters related to the performance and power consumption of the schedulable resource will also be different. Taking the schedulable resource as a cabinet, service or CPU as an example, the parameters related to the performance of the schedulable resource include: CPU main frequency, CPU utilization, etc.; the resource parameters related to the power consumption of the schedulable resource include: schedulable resource The rated power of the CPU, the actual power of the CPU, etc. If the schedulable resource is a host or a cabinet, the resource parameters related to the power consumption of the schedulable resource may also include: the rated power and actual power of the host or the cabinet.

Among them, the bidding parameters may include any parameters related to the bidding of schedulable resources, for example, may include the bid value of the schedulable resource, the benchmark bid value, the actual bid value, the time for participating in the bidding, the number of bidding times, and so on.

Based on the status information of the schedulable resources participating in the resource bidding in the resource market, the schedulable resources participating in the resource bidding can be regulated. Optionally, the regulation of schedulable resources participating in resource bidding includes at least one of the following regulation operations:

Regulation operation 1: According to the state information of the schedulable resources participating in the resource bidding in the resource market, adjust the actual competitive value of the schedulable resources participating in the resource bidding in the resource market.

Control operation 2: According to the state information of the schedulable resources participating in the resource bidding in the resource market, adjust the number of schedulable resources participating in the resource bidding.

Regarding regulation operation 1:

In some optional embodiments, in the regulation operation 1, the first schedulable resource whose bid value needs to be adjusted can be determined according to the state information of the schedulable resources participating in the resource bidding in the resource market; The resource bidding request, the first resource bidding request includes the first actual bid value, which is used to adjust the actual bid value of the first schedulable resource in the resource market.

Among them, the first schedulable resource is a schedulable resource for which it is necessary to adjust its competitive value in the resource market. For example, it can be a schedulable resource with a high competitive value, or a schedulable resource with a low competitive value, or a bidding. The schedulable resource whose value exceeds the set range, etc. Wherein, the first schedulable resource may be one or multiple.

Further optionally, the benchmark bid value and the transaction bid value of the schedulable resource can be combined to measure whether it is necessary to adjust the bid value of the schedulable resource in the resource market. Transaction bid value refers to the bid value related to the successful application of the schedulable resource in the resource market by the application. For example, it can be the actual bid value successfully applied for by a certain application in a certain time (such as the most recent time), or it can also be It is the average value of the actual competitive value of the schedulable resource that has been successfully applied for multiple times over a period of time. For schedulable resources, its transaction value is dynamically changing.

The benchmark bid value refers to the lowest bid value that a schedulable resource should meet to participate in bidding in the resource market. In this embodiment, different methods are allowed to define the benchmark competition value of schedulable resources. In an optional embodiment, the benchmark bid value of the schedulable resource can be determined based on the resource parameter that characterizes the operating state of the schedulable resource participating in the resource bidding in the resource market, and the benchmark bid value can reflect the performance of the schedulable resource. Consumption. In different time periods, the resource parameters that characterize the operating status of the schedulable resources participating in the resource bidding in the resource market will be different, which means that the benchmark bid value of the schedulable resources participating in the resource bidding will dynamically change.

Based on the above, the benchmark bid value of the schedulable resources participating in the resource bid can be calculated according to the resource parameters that characterize the operating status of the schedulable resources participating in the resource bidding in the resource market; from the characterization that the schedulable resources participating in the resource bidding are in the resource market In the bidding parameters of the bidding state, the transaction bid value of the schedulable resource participating in the resource bid is obtained; according to the benchmark bid value and the transaction bid value of the schedulable resource participating in the resource bidding, the first schedulable resource that needs to adjust the bid value is determined.

Further optionally, in the process of calculating the benchmark bid value of the schedulable resources participating in the resource bidding, a machine learning model, such as a cost prediction model, can be combined to characterize the operation status of the schedulable resources participating in the resource bidding in the resource market. The resource parameters are input to the cost prediction model, and the cost prediction model is used to calculate the benchmark bid value of schedulable resources participating in resource bidding. Using machine learning models can improve the accuracy and precision of the calculation results.

Further optionally, according to the benchmark bid value and the transaction bid value of the schedulable resources participating in the resource bidding, an implementation manner for determining the first schedulable resource whose bid value needs to be adjusted includes: setting the benchmark of the schedulable resources participating in the resource bidding Comparing the bid value and the transaction value; from the schedulable resources participating in the resource bidding, obtain the schedulable resource whose transaction value is less than the benchmark bid value as the first schedulable resource. If the transaction bid value of the schedulable resource is less than its benchmark bid value, it indicates that the bid value of the schedulable resource is too low. The cost of using this type of schedulable resource is relatively high, and it is necessary to increase its competitive value in the resource market. After that, the resource regulator 104 may send a first resource bid request to the resource scheduler 101, and the first resource bid request carries a first actual bid value greater than the reference bid value to participate in the bid of the first schedulable resource. Since the first actual bid value is higher than the benchmark bid value, if other applications need to use the first schedulable resource, they need to give a higher bid value than the first actual bid value, so as to improve the first schedulable resource in the resource market. The purpose of the competitive value in.

The foregoing selection of a schedulable resource whose transaction bid value is less than the reference bid value as the first schedulable resource is only an exemplary implementation, and is not limited to the foregoing example. For example, a schedulable resource whose transaction bid value is less than the benchmark bid value and the difference between the benchmark bid value and the transaction bid value is greater than the set difference threshold may also be selected as the first schedulable resource. For another example, a schedulable resource with a transaction bid value greater than the benchmark bid value and a difference greater than the set difference threshold (the bid value is too high) may also be selected as the first schedulable resource.

Regarding control operation 2:

In some optional embodiments, in the regulation operation 2, the second schedulable resource that needs to withdraw from the resource bid can be determined according to the state information of the schedulable resource participating in the resource bidding in the resource market; the second schedulable resource that needs to withdraw from the resource bidding can be determined; the second schedulable resource is sent to the resource scheduling party Resource bidding request. The second resource bidding request contains a second actual bid value for applying for a second schedulable resource; after applying for the second schedulable resource, adjust the working status of the second schedulable resource to the extent that it is not sufficient to be able to participate in the resource The bid request is to withdraw from the resource bidding. Among them, the second actual competitive value should be a competitive value that can ensure to a large extent that the resource control party 104 successfully applies for the second schedulable resource. For example, under the higher bidder mechanism, the second actual bid value should be a higher bid value; under the lower bidder mechanism, the second actual bid value should be a lower bid value, such as lower than The benchmark bid value of the second actual bid value.

Optionally, in an optional embodiment of determining the second schedulable resource that needs to withdraw from the resource bidding, the status information of the schedulable resource participating in the resource bidding may be analyzed based on the status information of the schedulable resource participating in the resource bidding. Load: From the schedulable resources participating in the resource bidding, obtain the schedulable resource whose load is lower than the first load threshold as the second schedulable resource. The load is low, indicating that the utilization rate of schedulable resources is not high, and its power consumption is relatively large. Withdrawing it from resource bidding can save this part of power consumption and stimulate applications to bid for schedulable resources with higher utilization rates. The utilization rate of other schedulable resources can be improved, which is conducive to reducing the overall power consumption.

In the foregoing optional embodiment, the load situation of a certain schedulable resource may be considered separately, and in addition, the overall load situation of the schedulable resource managed by the system may also be combined. If the overall load of schedulable resources is relatively light, it means that some schedulable resources will be withdrawn from resource bidding and will not affect system performance. Based on this, in another alternative embodiment, the load of the schedulable resource participating in the resource bidding may be analyzed according to the state information of the schedulable resource participating in the resource bidding in the resource market; it is determined that the load is lower than the second load. Whether the number of schedulable resources of the threshold is greater than the first number threshold; if the judgment result is yes, obtain the schedulable resource whose load is lower than the first load threshold from the schedulable resources participating in the resource bidding as the second schedulable resource ; Wherein, the second load threshold is greater than the first load threshold. If the number of schedulable resources whose load is lower than the second load threshold is greater than the first number threshold, it means that most of the schedulable resources have lighter load, and the schedulable resources with lighter load can be withdrawn from the resource bidding. It can save power consumption without affecting system performance.

In the above optional embodiment, the working status requirements for being able to participate in resource bidding can be preset, and by adjusting the working status of the second schedulable resource to not meet the requirements for being able to participate in resource bidding, it is possible to withdraw it from resource bidding. purpose. According to application requirements, the working status requirements that can participate in resource bidding can be flexibly set. Several examples are given below:

Example 1: The working state where schedulable resources can participate in resource bidding is the on state. Correspondingly, the second schedulable resource can be closed to withdraw it from the resource bidding.

Example 2: The working state of a schedulable resource that can participate in resource bidding is that it must be in a low power consumption mode higher than a specified level. Correspondingly, the second schedulable resource is controlled to enter a low power consumption mode not higher than the specified level, so as to withdraw it from the resource bidding.

Example 3: The working state of a schedulable resource that can participate in resource bidding is that its working frequency is higher than the specified frequency. Correspondingly, the working frequency of the second schedulable resource is adjusted to be no higher than the specified frequency, so that it is withdrawn from the resource bidding. Among them, the designated level or designated frequency corresponds to not meeting the requirement of being able to participate in resource bidding.

In other optional embodiments, in the control operation 2, it is also possible to determine the need to increase the schedulable resources that can participate in the resource bid based on the status information of the schedulable resources participating in the resource bidding in the resource market; never participate in the resource bidding Among the schedulable resources in, select the third schedulable resource that can participate in the resource bidding; and adjust the working state of the third schedulable resource to meet the requirements of being able to participate in the resource bidding to participate in the resource bidding.

Optionally, according to the state information of the schedulable resources participating in the resource bidding in the resource market, the load of the schedulable resources participating in the resource bid can be analyzed; if the load is higher than the third load threshold, the number of schedulable resources is greater than The second quantity threshold determines that it is necessary to increase the schedulable resources that can participate in the resource bidding. If the number of schedulable resources whose load is higher than the third threshold is greater than the second threshold, it means that most of the schedulable resources have a heavy load, and the schedulable resources that can participate in resource bidding may not be enough, and the system performance May be adversely affected, so some schedulable resources can be added to participate in resource bidding to ensure system performance.

Optionally, among the schedulable resources that have never participated in the resource bidding, the process of selecting the third schedulable resource that can participate in the resource bidding can be combined with a machine learning model, such as a screening model, using the screening model for those that have never participated in the resource bidding. The third schedulable resource that can participate in resource bidding is selected from the schedulable resources. Using machine learning models can improve selection efficiency and accuracy.

Among them, adjusting the working state of the third schedulable resource to meet the requirement of being able to participate in resource bidding is related to the requirement of being able to participate in resource bidding. For example, if the working state in which the schedulable resource can participate in the resource bidding is the on state, the third schedulable resource can be turned on to make it participate in the resource bidding. For another example, if the working state of a schedulable resource that can participate in resource bidding must be in a low-power mode higher than a specified level, the working state of the third schedulable resource can be adjusted to a low-power mode higher than the specified level. So as to make it participate in resource bidding. For another example, if the working state of the schedulable resource capable of participating in resource bidding is that its working frequency is higher than the specified frequency, the working frequency of the third schedulable resource can be adjusted to be higher than the specified frequency, so that it can participate in the resource bidding.

It should be noted that the foregoing embodiments are only exemplary implementations in which the resource regulator 104 uses a resource bidding mechanism to regulate schedulable resources participating in resource bidding, and the foregoing exemplary embodiments do not constitute a limitation on the protection scope of the present application.

Resource application and scheduling based on resource bidding mechanism:

In this embodiment, the resource demander 102 corresponding to the first application is taken as an example to describe the resource application and resource scheduling process based on the resource bidding mechanism. The first application program may be any application program that has resource requirements and has the authority to use schedulable resources managed by the resource management system 100.

For the resource demander 102, a resource bidding request may be generated for the first application program, and the resource bidding request may be sent to the resource scheduler 101 to participate in the resource bidding. For the resource scheduler 101, the resource bidding request sent by the resource demanding party 102 can be received, and according to the resource bidding request, among the schedulable resources participating in the resource bidding, the first application and other applications can apply for schedulable resources.

Further, the resource scheduler 101 may also return the resource application result of the first application to the resource demander 102. After receiving the resource application result of the first application, the resource demander 102 may also perform subsequent operations according to the resource application result.

It should be noted that there are multiple ways for the resource demander 102 to generate a resource bid request for the first application, and the information carried in the resource bid request may be different in different generation methods. Correspondingly, due to the different information carried in the resource bidding request, the resource scheduler 101 has different implementation manners for applying for schedulable resources for the first application and other applications among the schedulable resources participating in the resource bidding. Several exemplary implementations are given below:

Exemplary embodiment A1:

The resource demander 102 may generate a resource bid request in response to the resource application request. The resource bid request includes the first actual bid value given by the first application program for the requested resource; send the resource bid request to the resource dispatcher 101 to The resource scheduler 101 bids for the first application and other applications to apply for the target schedulable resource based on the first actual bid value; and receives the result of the first application's application for the target schedulable resource returned by the resource scheduler 101. The target schedulable resource is one or more schedulable resources among at least one schedulable resource managed by the resource management system 100.

For the resource scheduler 101, the resource bidding request of the first application program sent by the resource demander 102 can be received, and the first actual bid value given by the first application program for the requested resource can be obtained from the resource bidding request. In addition to the resource bidding request of the first application, the resource scheduler 101 will also receive resource bidding requests of other applications. Among these applications, some applications may apply for the same schedulable resources as the first application, thereby forming a bidding relationship with the first application. For ease of description, the schedulable resources applied for by the first application and other applications that have a bidding relationship with the first application are referred to as target schedulable resources. For the resource scheduler 101, it is necessary to determine other applications that compete with the first application to apply for the target schedulable resource and the actual bid value given by the other applications for the requested resource; then, according to the first application for the requested resource The given first actual bid value and the actual bid value given by other applications that have a bidding relationship with the first application for the application resource, determine whether the first application can apply for the target schedulable resource; and apply the first application The application result of the program for the target schedulable resource is returned to the resource demander 102.

Among them, the first actual bid value represents the actual price paid by the first application for the requested resource; correspondingly, other actual bid values represent the actual price paid by other applications for the requested resource. In this embodiment, the realization form of the first actual bid value or other actual bid values is not limited. For example, it may be virtual funds, actual funds, credit values, and points in the resource account corresponding to the first application or other applications. , Energy value or equity value, etc.

In the embodiment of the present application, the resource bidding request may include some other information in addition to the first actual bid value given by the first application for the requested resource. For example, the resource bidding request may also include the identification of the first application, and information capable of identifying the resources required by the first application. Wherein, the information capable of identifying the resource required by the first application may be the type of resource required by the first application, or may also be an identifier of the resource required by the first application. Of course, the information capable of identifying the resource required by the first application may also include the resource type required by the first application and the identifier of the resource required by the first application at the same time.

In the case that the resource bidding request includes information capable of identifying the resources required by the first application, the resource scheduler 101 may, according to the information capable of identifying the resources required by the first application included in the resource bidding request, in the resource bidding request Among schedulable resources, determine the target schedulable resource that the first application needs to apply for; then, based on the target schedulable resource, determine other applications that need to apply for the target schedulable resource, that is, there is a bidding relationship with the first application s application.

In the above example, in the process of determining whether the first application can apply for the target schedulable resource, the mechanism of "the higher the price gets" is adopted, that is, the resource scheduler 101 will use the first application as the first application given by the application resource. 1. The actual bid value is compared with the actual bid value given by other applications that have a bidding relationship with the first application for the application resource; if the first actual bid value is the highest actual bid value, it is determined that the first application can apply for the target. Scheduling resources; if the first actual bid value is not the highest actual bid value, it is determined that the first application fails to apply for the target schedulable resource. In addition to the "highest price" mechanism, other mechanisms such as "the lowest price" or "the closest average value" can also be used.

In the case of the "lowest price" mechanism, the resource scheduler 101 can compare the first actual bid value given by the first application for the requested resource with other applications (referring to the bidding relationship with the first application). Application) compare the actual competitive value given by the application resource; if the first actual competitive value is the lowest actual competitive value, determine that the first application applies for the target schedulable resource; if the first actual competitive value is not the lowest actual competitive value , It is determined that the first application fails to apply for the target schedulable resource.

In the case of adopting the mechanism of "the one who is closest to the average value", the resource scheduler 101 can calculate the first actual bid value given by the first application for the requested resource and other applications (referring to the existence of a bid with the first application). The relational application) is the average bid value of the actual bid value given by the application resource; compare the actual bid value given by the first actual bid and other applications for the applied resource with the average bid value; if the first actual bid The bid value is the bid value closest to the average bid value, and it is determined that the first application has applied for the target schedulable resource; if the first actual bid value is not the closest to the average bid value, it is determined that the first application has failed to apply To the target schedulable resources.

In the above-mentioned embodiment of this application, resource scheduling is combined with a resource bidding mechanism. The resource demander can give a reasonable actual bid value to participate in resource bidding according to the resource demand situation, and the resource scheduling party can participate in resource bidding according to the actual value given by each resource demander. Resource scheduling with competitive value is conducive to more reasonable allocation of resources and is conducive to improving overall resource utilization.

Exemplary embodiment A2:

The resource demander 102 determines at least one task that needs to be completed by the first application program and the resource information required by the at least one task in response to the resource application demand; and determines the resource required for the at least one task application based on the resource information that needs to be used by the at least one task The actual bid value given; based on the resource information that at least one task needs to use and the actual bid value given for the resource required for at least one task request, a resource bid request is generated; the resource bid request is sent to the resource scheduler 101 for The resource scheduler 101 bids for the first application and other applications to apply for schedulable resources; and receives the result of the first application's bid for schedulable resources returned by the resource scheduler 101.

Among them, the resource bid request includes the resource information that each task needs to use and the actual bid value corresponding to each task. The resource information that the task needs to use can be the type of resource that needs to be used, or the resource identifier.

For the resource scheduler 101, the resource bidding request of the first application program sent by the resource demanding party 102 can be received, and the resource information required by the at least one task and the actual bidding corresponding to each of the at least one task can be obtained from the resource bidding request. value. In addition to the resource bidding request of the first application, the resource scheduler 101 will also receive resource bidding requests of other applications. Among these resource bidding requests, there may be some application tasks that apply for the same schedulable resources as the tasks of the first application program, thereby forming a bidding relationship with the first application program. For the resource scheduler 101, a task that needs to be completed by the first application program can be used as a unit, and among the schedulable resources participating in the resource bidding, each task bids with other tasks to apply for schedulable resources. For any task that the first application needs to complete, the resource scheduler 101 can determine other tasks that apply for the same resource as the task and the actual competition value corresponding to other tasks; then, the actual competition value corresponding to the task corresponds to other tasks. To determine whether the task can apply for the required schedulable resources. After that, the resource scheduler 101 may feed back to the resource demander 102 the resource application result of each task to be completed by the first application program.

Optionally, the resource scheduler 101 may, after obtaining the resource application result of each task that needs to be completed by the first application, feed back the resource application result of each task to the resource demander 102; or, it may also obtain each task every time. The result of the resource application is fed back to the resource demander 102.

In an optional embodiment, the resource demander 102 may combine a machine learning model, such as an operation rule prediction model, to learn the operation rule of the first application program; according to the operation rule of the first application program, determine what the first application program needs to complete At least one task and at least one task needs to use resource information. Of course, in addition to using the machine learning model, it is also possible to analyze the operating law of the first application program manually or with the aid of a certain analysis tool. Whether it is manually, by means of analysis tools, or by using a machine learning model, the operation law of the first application program can be analyzed in combination with the log data generated by the first application program at the same period of history or within a specified time period. Through the running law of the first application program, it is possible to know: during which time period the first application program runs, which tasks need to be processed, which resources are used by these tasks, and the performance requirements of these tasks.

In an optional embodiment, in the process of determining the actual competitive value given for the resource required for at least one task application, the reference competitive value of the schedulable resource may be combined. Among them, the reference bid value of the schedulable resource can be its current benchmark bid value, current actual bid value or historically highest actual bid value, etc. As time goes by, the resource parameters of schedulable resources will change, and their benchmark competitive value will also change accordingly. Based on this, the candidate schedulable resource and its reference bid value can be determined from the known schedulable resource information according to the resource information that at least one task needs to use; according to the total budget price of this resource application and the reference bidding price of the candidate schedulable resource Value, determined as the actual competitive value given by the resources required for at least one task application. Among them, the candidate schedulable resource refers to the schedulable resource that at least one task is likely to apply for. Using the relative competitive value of these schedulable resources as a reference, the actual resource required for at least one task application can be given more accurately and reasonably. Competitive value. Wherein, the known schedulable resource information includes the reference bid value of the schedulable resource participating in the resource bidding, and this information can be transparently transmitted to the resource demander 102 by the market scheduler 101 or the resource regulator 104.

In an optional embodiment, the system 100 further includes: a resource management and control party 103. The resource management and control party 103 can provide schedulable resources to the system, and can also perform various management and control on the schedulable resources, such as power consumption control. Managed schedulable resources mainly refer to schedulable resources that are subjected to power consumption control operations. The running status of these schedulable resources will change. These changes will affect the number of schedulable resources participating in resource bidding, which can indirectly affect resource demand. The actual competitive value given by the party 102 for each task. Based on this, after applying for the required resources for at least one task and giving the actual competitive value, the resource demander 102 can also use the resource management and control party 103 to feed back the operation status information of the managed and schedulable resources in the future period, which is at least one The actual competitive value given by the resources required for the task application will be adjusted.

For example, it can be determined whether the number of resources required for at least one task is less than the third number threshold in the future period according to the running status information of the managed and schedulable resources fed back by the resource management and control party; if the result of the judgment is yes, It means that the amount of resources required for at least one task will be relatively scarce. In order to increase the probability of successfully applying for the required resources, the actual competitive value given for the resources required for at least one task application can be increased.

Further, after the actual bid value is given for the resources required for the at least one task application, a resource bid request may be generated based on the resource information required for the at least one task and the actual bid value given for the resources required for the at least one task application.

In an alternative embodiment, resource bidding requests may be generated for each task separately, and the resource bidding request for each task includes the resource information that the task needs to use and the actual bid value given for the required resources for the task application. Or, it is also possible to encapsulate the resource information required for at least one task and the actual bid value given for the resource required for the at least one task application into the same resource bidding request.

Regardless of the implementation mode, the resource scheduler 101 can obtain the resource information that each task needs to use and the actual competitive value given for the required resources for each task application, and then the task can be used as a unit. Bidding resources for each task.

In this exemplary embodiment, the resource demander 102 determines the tasks that the first application program needs to complete, and determines the resource information it needs in units of tasks. This is conducive to selecting more reasonable resources for the first application program. Reduce unnecessary waste of resources and save power consumption.

Exemplary embodiment A3:

In response to the resource application demand, the resource demander 102 determines the number of resources that the first application needs to apply for this time and the total budget price of this resource application; based on the number of resources that need to be applied this time and the total budget price of this resource application, Generate a resource bid request; send the resource bid request to the resource scheduler 101 for the resource scheduler 101 to bid for the first application with other applications for schedulable resources; receive the first application bid request returned by the resource scheduler 101 The result of schedulable resources.

Among them, the resource bidding request includes the number of resources that the first application needs to apply for this time and the total budget price of this resource application.

Optionally, the resource demander 102 may determine the at least one task that the first application program needs to complete and the resource information that the at least one task needs to use according to the running law of the first application program; determine the resource information that needs to be used by the at least one task The number of resources that the first application needs to apply for this time.

In an optional embodiment, the resource demander 102 may combine a machine learning model, such as an operation rule prediction model, to learn the operation rule of the first application program; according to the operation rule of the first application program, determine what the first application program needs to complete At least one task and at least one task needs to use resource information. For related descriptions of the machine learning model, please refer to the foregoing embodiments, which will not be repeated here.

For the resource scheduler 101, a resource bidding request can be received. The resource bidding request includes: the number of resources that the first application needs to apply for this time and the total budget of this resource application; according to the first application's current needs From the schedulable resources participating in the resource bidding, determine at least one resource to be applied for and its actual bid value; based on the actual bid value of at least one resource to be applied for, it is the first application The program bids with other applications to apply for at least one resource to be applied for, and returns the resource application result of the first application.

In an optional embodiment, the resource scheduler 101 may determine at least one resource to be applied for among the schedulable resources participating in resource bidding according to the number of resources to be applied for this time; according to the total budget price of this resource application and at least The reference bid value of a resource to be applied for is to determine the actual bid value of at least one resource to be applied for.

Among them, the reference competitive value of the resource to be applied for includes, but is not limited to: the current benchmark competitive value of the resource to be applied for, the maximum actual competitive value in the resource market, or the average competitive value in the resource market.

In an optional embodiment, the resource scheduler 101 may determine at least one resource to be applied for among the schedulable resources participating in resource bidding according to the number of resources to be applied for this time; according to the number of resources to be applied for this time and the number of resources that need to be applied for this time The total budgeted price of resource applications is calculated and the average bid value is calculated as the actual bid value of at least one resource to be applied for.

After obtaining the actual bid value of each resource to be applied for, the resource dispatcher can bid for each resource to be applied for for the first application program and other applications according to the actual bid value of each resource to be applied for. In the bidding application process, the higher bidder’s mechanism, the lower bidder’s mechanism or the nearest average taker mechanism, etc. can be used. For detailed bidding processes under different mechanisms, please refer to the description in the foregoing embodiment. Go into details again.

In this embodiment, the resource demander 102 only needs to provide the resource scheduler 101 with the amount of resources and the total budget required by the first application. Other operations are performed by the resource scheduler 101, and the resource demander 102 has a relatively large burden. Light and easy to implement.

It should be noted that in each embodiment of the present application, power consumption management and control, resource bidding, and regulation mechanisms can be combined. As the power consumption control party, the resource regulator 104 can combine the overall power consumption of at least one schedulable resource. With the goal of reducing overall power consumption, the schedulable resources participating in resource bidding are regulated. On the one hand, it can improve the rationality of resource scheduling, on the other hand, it can reduce overall power consumption and improve overall energy efficiency. Among them, there are many ways to combine power consumption control, resource bidding, and regulation mechanisms. The following embodiment provides a combination method.

In this embodiment, the competitive value of the schedulable resource is combined with the power consumption of the schedulable resource. For example, the resource scheduler 104 considers the power consumption cost of the schedulable resource in the process of calculating the benchmark bid value of the schedulable resource; further, in the process of scheduling the schedulable resource participating in the resource bidding, the schedulable resource Based on the benchmark bid value, this is equivalent to scheduling the schedulable resources participating in the resource bidding on the basis of the power consumption cost of the schedulable resource. In addition, the actual competitive value given by each application for application resources either refers to or is affected by the benchmark competitive value of schedulable resources. Therefore, the actual competitive value given by each application also reflects the schedulable resource’s benchmark competitive value. The power consumption cost of scheduling resources.

Among them, factors that have an impact on the power consumption cost of the schedulable resource include: the power consumption of the schedulable resource itself. Among them, the resource parameters (such as rated power, actual power, etc.) of the schedulable resource can reflect the power consumption of the schedulable resource to a certain extent. The resource regulator 104 may calculate the power consumption of the schedulable resource according to the resource parameter of the schedulable resource. In addition, factors that have an impact on the power consumption cost of schedulable resources include many other factors. In this embodiment, the main consideration is the power consumption of the infrastructure on which the schedulable resource depends. Infrastructure refers to the facilities that provide basic services for schedulable resources. According to the different schedulable resources, the infrastructure will be different, and there is no restriction on this. Take the server as an example of a schedulable resource. The infrastructure can include cabinets that provide support services for the servers, air-conditioning equipment, fans or water-cooling equipment for cooling the servers, power equipment for powering the servers, switches that provide network connections for the servers, etc. .

Based on the above, the resource control party 104 calculates the benchmark bid value of the schedulable resource according to the resource parameter that characterizes the operating state of the schedulable resource in the resource market. One embodiment includes: according to the operating state of the schedulable resource in the resource market. Calculate the power consumption of the schedulable resource; predict the power consumption cost of the schedulable resource according to the power consumption data of the schedulable resource and the infrastructure on which the schedulable resource depends; determine the power consumption cost of the schedulable resource The benchmark competition value of scheduling resources. Among them, the power consumption of the infrastructure on which the schedulable resource depends is converted into the power consumption cost of the schedulable resource, and the measurement of the power consumption cost of the schedulable resource will be more reasonable, which is conducive to improving the rationality of resource scheduling.

Among them, the greater the power consumption cost of the schedulable resource, the higher the contribution of the schedulable resource to the system energy consumption, the higher the power consumption cost, and the higher the benchmark competition value; conversely, the higher the power consumption cost of the schedulable resource Small means that the schedulable resource has a relatively low contribution to the system's energy consumption. The lower the power consumption cost, the lower the benchmark competition value.

The following describes the effects brought about by the technical solution of the present application in combination with two examples in the IDC scenario:

The first example: Through reasonable resource regulation, the overall power consumption of IDC can be reduced

Assume that there are two cabinets in the IDC computer room shown in Figure 1e, cabinet 1 and cabinet 2, respectively. There are 4 servers running on cabinet 1 and 2 servers running on cabinet 2. In order to reduce the temperature of the server in the cabinet 1, the cooling system of the cabinet 1 needs 1200w for normal operation. Similarly, the cooling system of cabinet 2 needs 1000w for normal operation. The average power consumption of each server in the cabinet 1 is 1200w/4=300w, and the average power consumption of each server in the cabinet 2 is 1000w/2=500w.

On the one hand, the resource regulator 104 can bid for the servers in the cabinet 2 with a higher actual bid value, thereby increasing the actual competitive value of the servers in the cabinet 2 in the resource market. For the application program, it will consider buying cheaper resources rather than more expensive resources. Therefore, the application program can be prevented from purchasing the server resources in the cabinet 2, so that the server resources in the cabinet 2 can withdraw from the resource bidding and save the overall power consumption.

On the other hand, the resource control party 104 can reduce the actual competitive value of the servers in the cabinet 1 to attract more applications to purchase the server resources in the cabinet 1 so as to give full play to the advantages of low-power devices and save overall power consumption.

The second example: reduce the power consumption of the server through reasonable resource scheduling

Assuming that an application requires a server with a performance of 100 to run for 1 hour, the total budget that the application can control is 1150, and there are currently three servers, namely Server 1 (performance: 95, competitive value: 850), Server 2 (performance: 90, competitive value: 700) and server 3 (performance: 110, competitive value: 1200). Among them, the competitive value represents the power consumption cost of the server; the high competitive value represents the high power consumption cost; the low competitive value represents the low power consumption cost.

According to the traditional resource scheduling scheme, resource 3 will be directly allocated to the application, and the actual power consumption cost is 1200*1 hour. However, in this embodiment, based on the resource bidding method, the resource demander 102 or the resource scheduler 104 can perceive the power consumption costs of different servers, and can allocate resources more reasonably. For example, the resource demander 102 or the resource scheduler 104 can purchase resources for 2 and a half hours according to the task that the application needs to complete, and then purchase resources for 3 and a half hours, so that the overall power consumption cost is half an hour of the power consumption of resource 2 plus 350 The power consumption of resource 3 in the first half hour is 600, which is 950 in total. Compared with the power consumption of 1200, the scheduling method of this embodiment saves 250, but the resource performance obtained by the user is not lost, but the actual power consumption is reduced.

It is explained here that the resource scheduler 101, the resource demander 102, the resource regulator 104, and the resource management controller 103 in the resource scheduling system 100 described above correspond to the resource scheduling node, application layer, and application layer in the system embodiment shown in FIG. 1b, respectively. Resource control node and resource manager. For a detailed description of the resource scheduling node, application layer, resource control node, and resource manager, refer to the description of the resource scheduling party 101, the resource demanding party 102, the resource regulating party 104, and the resource managing party 103 in the embodiment shown in FIG. 1d. The description will not be repeated here.

In terms of deployment and implementation, the resource management system 100 provided in the above-mentioned embodiment of the present application can adopt a centralized deployment mode, that is, the resource scheduler 101, the resource demander 102, the resource management and control party 103, and the resource control party 104 are deployed on the same physical device (e.g., Servers, laptops or desktops), virtual machines or containers; alternatively, a distributed deployment method can also be adopted, that is, the resource scheduler 101, the resource demander 102, the resource control party 103, and the resource control party 104 are deployed in multiple locations. It can be implemented in a physical device, virtual machine, or container. For example, the resource scheduling party 101, the resource demanding party 102, the resource management and control party 103, and the resource control party 104 are respectively deployed in four different servers. For another example, the resource scheduling party 101 and the resource demanding party 102 are deployed in one server, and the resource management and control party 103 and the resource control party 104 are deployed in another server. Of course, the resource scheduling party 101, the resource demanding party 102, the resource management and control party 103, and the resource control party 104 can also be deployed in the cloud as required.

The resource management system 100 provided in this embodiment can be deployed in any application scenario with resource scheduling and control requirements, for example, it can be deployed in a computer room system, or can be deployed in a data center system. The following embodiments will respectively provide a computer room system and a data center system for resource management based on a market competition mechanism.

FIG. 2 is a schematic diagram of the structure and working principle of a computer room system provided by an exemplary embodiment of this application. As shown in FIG. 2, the computer room system 200 of this embodiment includes: a computer room. The computer room refers to a physical place where machinery and equipment are stored, such as a room or a factory building. The computer room system 200 can be constructed separately or placed in other buildings. Further, as shown in FIG. 2, the computer room system 200 further includes: at least one physical device 201, a resource scheduling device 202, a resource application device 203, and a resource control device 204 located in the computer room. This embodiment does not limit the number of physical devices 201, and it may be one or multiple.

In this embodiment, the device form of the physical device 201 is not limited. As shown in FIG. 2, the physical device 201 may be an IT-type device in a computer room and a refrigeration device for cooling the IT-type device, such as an air-conditioning device. For example, the at least one physical device 201 may include, but is not limited to: cabinet equipment, server equipment, terminal equipment, printers, hubs, power supply equipment, storage equipment, network switching equipment, air conditioning equipment, water cooling equipment, and so on. The server device may include, but is not limited to, a conventional server, a server array, or a cloud server. The power supply equipment can be storage battery equipment, dry battery equipment, or uninterruptible power supply (UPS). Storage devices may include, but are not limited to: disks, disk arrays, hard disks, network storage devices (NAS), and so on.

Similarly, this embodiment does not limit the device forms of the resource scheduling device 202, the resource application device 203, and the resource control device 204. The resource scheduling device 202, the resource application device 203, and the resource control device 204 can be any computer devices with certain computing and processing capabilities, such as smart phones, laptops, desktop computers, tablet computers, smart watches, and other terminal devices, or they can be Server equipment such as conventional servers, cloud servers, or server arrays.

In this embodiment, from the perspective of resource scheduling and regulation, the physical device 201 in the computer room is divided into at least one schedulable resource and at least one infrastructure. Among them, schedulable resources refer to resource devices, modules, or units that can provide certain services, functions, or resources (for example, computing, storage, or network) to the outside in the computer room. For example, the schedulable resource can be a physical device such as a server or printer in a computer room, or a hardware module such as a CPU, GPU, memory, network card, or hard disk on the server, or a virtual machine running on a physical device, Software modules or units such as cloud services. Generally speaking, physical servers are the most common schedulable resource. Infrastructure refers to the facilities in the computer room that provide basic services for schedulable resources. According to the different schedulable resources, the infrastructure will be different, and there is no restriction on this. Take the server as an example of a schedulable resource. The infrastructure can include cabinets that provide support services for the servers, air-conditioning equipment, fans or water-cooling equipment for cooling the servers, power equipment for powering the servers, switches that provide network connections for the servers, etc. .

Among them, schedulable resources are available resources in the computer room system 200, and the schedulable resources can be reasonably scheduled and regulated through the resource scheduling device 202 and the resource control device 204, which is beneficial to improve the resource utilization efficiency of the computer room system 200 and save energy. , It is of great significance to improve resource sharing and reduce operating costs.

Among them, the resource application device 203 mainly provides resource application services for at least one application program. There may be one or more resource application devices 203. For example, the resource application device 203 may run services, programs, plug-ins, patches, etc. related to the resource application, so as to realize the provision of resource application services for the application. In this embodiment, the resource application device 203 and the resource scheduling device 202 cooperate with each other to realize a resource scheduling scheme based on a resource bidding mechanism, which can make resource scheduling more reasonable and help improve overall resource utilization.

Wherein, the resource application device 203 and the resource scheduling device 202 cooperate with each other, and the process of resource scheduling for any application program with resource requirements is the same or similar. In this embodiment, taking the application of the first application program for schedulable resources as an example, the process of resource application and resource scheduling based on the resource bidding mechanism is described. The first application program may be any application program that has permission to use at least one schedulable resource.

For the resource application device 203, a resource bidding request may be generated for the first application program, and the resource bidding request may be sent to the resource scheduling device 202 to participate in the resource bidding. For the resource scheduling device 202, the resource bidding request sent by the resource applying device 203 can be received, and according to the resource bidding request, among the schedulable resources participating in the resource bidding, apply for schedulable resources for the first application and other applications.

For the resource control device 204, it can monitor the status information of the schedulable resources participating in the resource bidding in the resource market; according to the status information of the schedulable resources participating in the resource bidding in the resource market, perform the monitoring of the schedulable resources participating in the resource bidding Regulation.

Regarding the cooperation of the resource scheduling device 202, the resource application device 203, and the resource control device 204, the detailed process of resource scheduling and control based on the resource bidding and control mechanism and other related content can be referred to the foregoing embodiment, and will not be repeated here.

In the data center system of this embodiment, resource scheduling, resource bidding, and control mechanisms are combined, and schedulable resources can be scheduled according to the resource requirements of the application program, which can allocate resources more reasonably, which is conducive to improving overall resources. Utilization rate. Further, in a scenario that combines power consumption control, resource bidding and regulation mechanisms, with the goal of reducing power consumption costs, resource scheduling based on resource bidding and regulation mechanisms is conducive to improving overall energy efficiency.

FIG. 3 is a schematic structural diagram of a data center system provided by an exemplary embodiment of this application. As shown in FIG. 3, the data center system 300 includes: at least one computer room 301, a resource scheduling device 302, a resource application device 303, and a resource control device 304. The data center system 300 can be built separately, or placed in other buildings, or distributed in different geographical locations. Among them, each computer room 301 includes at least one physical device 305. The number of physical devices contained in each computer room 301 may be one or multiple.

The computer room 301 in this embodiment is similar to the computer room in the embodiment shown in FIG. 2. For the related description of the physical device 305 and the computer room 301, please refer to the description in the embodiment shown in FIG. 2, which will not be repeated here.

In this embodiment, from the perspective of resource scheduling, the physical device 305 in each computer room 301 is divided into at least one schedulable resource and at least one infrastructure. Among them, schedulable resources refer to resource devices, modules, or units that can provide certain services, functions, or resources (for example, computing, storage, or network) to the outside in the computer room. For related descriptions of schedulable resources and infrastructure, please refer to the embodiment shown in FIG. 2, which will not be repeated here.

The difference between this embodiment and the embodiment shown in FIG. 2 is that the resource scheduling device 302, the resource application device 303, and the resource control device 304 no longer specifically serve a specific computer room, but serve the entire data center system 300, and need to The schedulable resources in each computer room 301 perform resource scheduling. Although the resource scheduling device 302 no longer specifically serves a specific computer room, it can be physically deployed in a certain computer room or independently deployed outside each computer room.

In the same way, the resource application device 303 of this embodiment mainly provides resource application services for application programs in each computer room. The resource application device 303 cooperates with the resource scheduling device 302 and the resource control device 304 to realize a resource scheduling and control scheme based on the resource bidding and control mechanism, which can make resource scheduling and control more rational and help improve the overall resource utilization rate. .

Regarding the cooperation of the resource application device 303, the resource scheduling device 302, and the resource control device 304, the detailed process and related content of resource scheduling and control for any application program with resource requirements can be referred to the foregoing embodiment, and will not be repeated here.

Fig. 4a is a schematic flowchart of a resource control method provided by an exemplary embodiment of this application. As shown in Figure 4a, the method includes:

41a. Monitor the status information of the schedulable resources participating in the resource bidding in the resource market.

42a. According to the state information of the schedulable resources participating in the resource bidding in the resource market, control the schedulable resources participating in the resource bidding.

The method in this embodiment is mainly described from the perspective of a resource control party, a resource control device, or a resource control node, and mainly describes how to perform resource control on schedulable resources participating in resource bidding according to a resource control mechanism.

In an optional embodiment, step 42a includes at least one of the following regulation operations:

According to the state information of the schedulable resources participating in the resource bidding in the resource market, adjust the actual competitive value of the schedulable resources participating in the resource bidding in the resource market;

According to the state information of the schedulable resources participating in the resource bidding in the resource market, the number of schedulable resources participating in the resource bidding is adjusted.

In an optional embodiment, according to the state information of the schedulable resources participating in the resource bidding in the resource market, adjusting the actual bid value of the schedulable resources participating in the resource bidding in the resource market includes: according to the schedulable resources participating in the resource bidding The status information of the resource in the resource market determines the first schedulable resource whose bid value needs to be adjusted; sends a first resource bid request to the resource scheduler, and the first resource bid request includes the first actual bid value to adjust the first schedulable resource. The actual competitive value of scheduling resources in the resource market.

Further, according to the state information of the schedulable resources participating in the resource bidding in the resource market, an implementation manner for determining the first schedulable resource that needs to adjust the bid value includes: according to the characterization that the schedulable resources participating in the resource bidding are in the resource market Calculate the benchmark bid value of the schedulable resources participating in the resource bidding; obtain the schedulable resources participating in the resource bidding from the bidding parameters representing the bidding status of the schedulable resources participating in the resource bidding in the resource market Transaction bid value: According to the benchmark bid value and the transaction bid value of the schedulable resources participating in the resource bidding, determine the first schedulable resource whose bid value needs to be adjusted.

Further optionally, an implementation manner for calculating the benchmark bid value of the schedulable resource participating in the resource bidding based on the resource parameter that characterizes the operating state of the schedulable resource participating in the resource bidding in the resource market includes: The resource parameters of the running state of the schedulable resources in the resource market are input into the cost prediction model, and the cost prediction model is used to calculate the benchmark bid value of the schedulable resources participating in the resource bidding.

Further optionally, according to the benchmark bid value and the transaction bid value of the schedulable resources participating in the resource bidding, an implementation manner for determining the first schedulable resource whose bid value needs to be adjusted includes: from among the schedulable resources participating in the resource bidding , Obtain the schedulable resource whose transaction bid value is less than the benchmark bid value as the first schedulable resource; among them, the first actual bid value is greater than the benchmark bid value.

In an optional embodiment, adjusting the number of schedulable resources participating in the resource bidding according to the status information of the schedulable resources participating in the resource bidding in the resource market includes: according to the schedulable resources participating in the resource bidding in the resource market Status information, determine the second schedulable resource that needs to withdraw from the resource bidding; send a second resource bidding request to the resource scheduler, the second resource bidding request includes the second actual bid value, used to apply for the second schedulable resource; After the second schedulable resource, the working state of the second schedulable resource is adjusted to not meet the requirement of being able to participate in the resource bidding to withdraw from the resource bidding.

Optionally, according to the state information of the schedulable resources participating in the resource bidding in the resource market, an implementation manner for determining the second schedulable resource that needs to withdraw from the resource bidding includes: according to the schedulable resources participating in the resource bidding in the resource market Analyze the load of the schedulable resource participating in the resource bidding; obtain the schedulable resource whose load is lower than the first load threshold from the schedulable resource participating in the resource bid as the second schedulable resource.

Furthermore, before acquiring the schedulable resource whose load is lower than the first load threshold as the second schedulable resource, the method further includes: determining whether the number of schedulable resources whose load is lower than the second load threshold is greater than the first quantity Threshold; if the judgment result is yes, from the schedulable resources participating in the resource bidding, obtain the schedulable resource whose load is lower than the first load threshold as the second schedulable resource; wherein, the second load threshold is greater than the first load The amount threshold.

Optionally, the foregoing adjustment of the working state of the second schedulable resource to the requirement of being able to participate in the resource bidding to withdraw from the resource bidding includes the following methods:

Close the second schedulable resource; or

Control the second schedulable resource to enter a low power consumption mode not higher than the specified level; or

Adjust the working frequency of the second schedulable resource to not higher than the specified frequency;

Among them, the designated level or designated frequency corresponds to not meeting the requirement of being able to participate in resource bidding.

In an optional embodiment, adjusting the number of schedulable resources participating in the resource bidding according to the status information of the schedulable resources participating in the resource bidding in the resource market includes: according to the schedulable resources participating in the resource bidding in the resource market Status information, determine the need to increase the schedulable resources that can participate in resource bidding; select the third schedulable resource that can participate in the resource bidding from the schedulable resources that have never participated in the resource bidding; and adjust the working status of the third schedulable resource to Meet the requirements to be able to participate in resource bidding to participate in resource bidding.

Further optionally, according to the state information of the schedulable resources participating in the resource bidding in the resource market, an implementation manner for determining the need to increase the schedulable resources that can participate in the resource bidding includes: according to the schedulable resources participating in the resource bidding in the resource State information in the market, analyze the load of schedulable resources participating in resource bidding; if the number of schedulable resources whose load is higher than the third load threshold is greater than the second threshold, determine the need to increase the schedulable resources that can participate in the resource bidding Resources.

Fig. 4b is a schematic flowchart of a resource application method provided by an exemplary embodiment of this application. As shown in Figure 4b, the method includes:

41b. Determine at least one task to be completed by the first application program and resource information to be used by the at least one task.

42b. Determine the actual competitive value given for the resources required by the at least one task application based on the information about the resources that at least one task needs to use.

43b. Generate a resource bid request based on the resource information that at least one task needs to use and the actual bid value given for the resource required by the at least one task application.

44b. Send the resource bidding request to the resource scheduler, so that the resource scheduler can bid for the first application and other applications to apply for schedulable resources.

45b. Receive the result of the first application bidding for schedulable resources returned by the resource scheduler.

The method in this embodiment is mainly described from the perspective of the resource demander, the resource application device or the application layer. It mainly describes how to generate a resource bid request for the first application, and based on the resource bid request, the first application sends the resource dispatcher to the resource dispatcher. , The resource scheduling device or resource scheduling node applies for schedulable resources.

In an optional embodiment, determining the at least one task to be completed by the first application program and the resource information that the at least one task needs to use includes: using an operation law prediction model to learn the operation law of the first application program; according to the first application The running law of the program determines at least one task to be completed by the first application program and resource information to be used by the at least one task.

In an optional embodiment, determining the actual competitive value given for the resources required by the at least one task application based on the resource information required for the at least one task includes: according to the resource information required for the at least one task, from the known available resources The candidate schedulable resource and its reference bid value are determined in the scheduling resource information; the actual bid value given by the resource required for at least one task application is determined according to the total budget price of the resource application and the reference bid value of the candidate schedulable resource.

In an optional embodiment, generating a resource bid request based on the resource information needed to be used by the at least one task and the actual bid value given for the resource required for the at least one task application includes: generating a resource bid request for each task, The resource bidding request for each task includes the resource information that the task needs to use and the actual bid value given for the resource required for the task application; or, the resource information that needs to be used for at least one task and the resource required for at least one task application. The actual bid value of is encapsulated in the same resource bid request.

In an optional embodiment, after determining the actual competitive value given for the resource required for the at least one task application, the method further includes: according to the operation status information of the managed schedulable resource fed back by the resource management and control party in the future period, Adjust the actual competitive value given for the resources required to apply for at least one task.

In an optional embodiment, according to the running status information of the managed and schedulable resources fed back by the resource management and control party in the future period, the adjustment of the actual competitive value given by the resource required for the at least one task application includes: according to the resource The running status information of the managed and schedulable resources fed back by the management and control party in the future period, to determine whether the number of resources required for at least one task is less than the third number threshold in the future period; if the result of the judgment is yes, increase to at least one The actual competitive value given by the resources required for the task application.

Fig. 4c is a schematic flowchart of another resource application method provided by an exemplary embodiment of this application. As shown in Figure 4c, the method includes:

41c. Determine the number of resources that the first application needs to apply for this time and the total budget for this resource application.

42c. Generate a resource bid request based on the number of resources to be applied for this time and the total budget price of this resource application.

43c. Send the resource bidding request to the resource scheduler, so that the resource scheduler can bid for the first application and other applications to apply for schedulable resources.

44c. Receive the result of the first application bidding for schedulable resources returned by the resource scheduler.

In an optional embodiment, determining the number of resources that the first application program needs to apply for this time includes: determining at least one task that the first application program needs to complete and at least one task that needs to be used according to the running law of the first application program Resource information: According to the resource information that at least one task needs to use, determine the number of resources that the first application needs to apply for this time.

Fig. 4d is a schematic flowchart of a resource scheduling method provided by an exemplary embodiment of this application. As shown in Figure 4d, the method includes:

41d. Receive a resource bidding request. The resource bidding request includes: the number of resources that the first application needs to apply for this time and the total budget price of this resource application.

42d. According to the number of resources to be applied for this time and the total budget price of this resource application, determine at least one resource to be applied for and its actual bid value among the schedulable resources participating in the resource bidding.

43d. According to the actual bid value of the at least one resource to be applied for, bid for at least one resource to be applied for for the first application and other applications, and return the resource application result of the first application.

The method in this embodiment is mainly described from the perspective of a resource scheduling party, a resource scheduling device, or a resource scheduling node, and mainly describes how to allocate resources to schedulable resources based on the resource bidding mechanism and in combination with the resource requirements of the application program.

In an optional embodiment, according to the number of resources to be applied for this time and the total budget price of this resource application, at least one resource to be applied for and its actual bid value are determined among the schedulable resources participating in the resource bidding, including: According to the number of resources to be applied for this time, at least one resource to be applied for is determined among the schedulable resources participating in resource bidding; according to the total budget price of this resource application and the reference bid value of at least one resource to be applied for, at least one resource to be applied for is determined The actual competitive value of the requested resource.

Optionally, the reference bid value of the at least one resource to be applied for includes: the benchmark bid value of the at least one resource to be applied for, the average actual bid value in the resource market, or the maximum actual bid value in the resource market.

In an optional embodiment, the actual competitive value of at least one resource to be applied for is determined according to the number of resources to be applied for this time and the total budget price of this resource application, including: according to the number of resources to be applied for this time, participating in Among the schedulable resources of resource bidding, at least one resource to be applied for is determined; based on the number of resources to be applied for this time and the total budgeted price of this resource application, the average bid value is calculated as the actual bid value of at least one resource to be applied for.

It should be noted that the execution subject of each step of the method provided in the foregoing embodiment may be the same device, or different devices may also be the execution subject of the method. For example, the execution subject of steps 41d to 44d may be device A; for another example, the execution subject of

steps

41d and 42d may be device A, and the execution subject of step 43d may be device B; and so on.

In addition, in some of the processes described in the above embodiments and drawings, multiple operations appearing in a specific order are included, but it should be clearly understood that these operations may be performed out of the order in which they appear in this document or performed in parallel. The operation sequence numbers such as 41a, 42a, etc., are only used to distinguish different operations, and the sequence numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions of "first" and "second" in this article are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, nor do they limit the "first" and "second" Are different types.

Fig. 5a is a schematic structural diagram of a resource control device provided by an exemplary embodiment of this application. As shown in FIG. 5a, the resource control device includes: a memory 51a and a processor 52a.

The memory 51a is used to store computer programs and can be configured to store various other data to support operations on the resource control device. Examples of these data include instructions for any application or method operating on the resource control device, contact data, phone book data, messages, pictures, videos, etc.

The processor 52a, coupled with the memory 51a, is used to execute a computer program for: monitoring the status information of the schedulable resources participating in the resource bidding in the resource market; according to the status information of the schedulable resources participating in the resource bidding in the resource market , Regulate and control the schedulable resources participating in resource bidding.

In an optional embodiment, when the processor 52a regulates the schedulable resources participating in the resource bidding, it is specifically configured to perform at least one of the following regulation operations:

In an optional embodiment, when the processor 52a adjusts the actual bid value of the schedulable resource participating in the resource bidding in the resource market, it is specifically configured to: according to the state information of the schedulable resource participating in the resource bidding in the resource market, Determine the first schedulable resource for which the bid value needs to be adjusted; send a first resource bid request to the resource dispatcher, the first resource bid request includes the first actual bid value to adjust the actual bid of the first schedulable resource in the resource market value.

Further, when the processor 52a determines the first schedulable resource for which the bid value needs to be adjusted, it is specifically configured to: calculate the schedulable resource participating in the resource bidding according to the resource parameter that characterizes the operating state of the schedulable resource participating in the resource bidding in the resource market The benchmark bid value of the resource; from the bidding parameters that characterize the bidding status of the schedulable resource participating in the resource bidding in the resource market, the transaction bid value of the schedulable resource participating in the resource bidding is obtained; according to the benchmark of the schedulable resource participating in the resource bidding The bid value and the transaction bid value determine the first schedulable resource that needs to adjust the bid value.

Further optionally, when the processor 52a calculates the benchmark bid value of the schedulable resource participating in the resource bidding, it is specifically configured to: input the resource parameter representing the operating state of the schedulable resource participating in the resource bidding in the resource market into the cost prediction model , Use the cost prediction model to calculate the benchmark bid value of schedulable resources participating in resource bidding.

Further optionally, when the processor 52a determines the first schedulable resource for which the bid value needs to be adjusted, it is specifically configured to obtain the schedulable resource whose transaction bid value is less than the reference bid value from the schedulable resources participating in the resource bidding as the first schedulable resource. 1. Dispatchable resources; among them, the first actual bid value is greater than the benchmark bid value.

In an optional embodiment, when the processor 52a adjusts the number of schedulable resources participating in the resource bidding, it is specifically configured to: according to the state information of the schedulable resources participating in the resource bidding in the resource market, determine the ones that need to withdraw from the resource bidding Second schedulable resource; send a second resource bidding request to the resource scheduler, the second resource bidding request includes the second actual bid value, used to apply for the second schedulable resource; after applying for the second schedulable resource, adjust the second resource bidding request 2. The working state of the schedulable resource cannot meet the requirement of being able to participate in the resource bidding to exit the resource bidding.

Optionally, when the processor 52a determines the second schedulable resource that needs to withdraw from the resource bidding, it is specifically configured to: analyze the status information of the schedulable resource participating in the resource bidding in the resource market according to the status information of the schedulable resource participating in the resource bidding. Load: From the schedulable resources participating in the resource bidding, obtain the schedulable resource whose load is lower than the first load threshold as the second schedulable resource.

Furthermore, the processor 52a is further configured to: determine whether the number of schedulable resources whose load is lower than the second load threshold is greater than the first number threshold; if the determination result is yes, obtain from the schedulable resources participating in the resource bidding The schedulable resource whose load is lower than the first load threshold is used as the second schedulable resource; wherein, the second load threshold is greater than the first load threshold.

Optionally, when the processor 52a adjusts the working state of the second schedulable resource to a requirement of being able to participate in resource bidding, it is specifically configured to:

Close the second schedulable resource; or

The working frequency of the second schedulable resource is adjusted to be no higher than the specified frequency; wherein, the specified level or specified frequency corresponds to not meeting the requirement of being able to participate in resource bidding.

In an alternative embodiment, when the processor 52a adjusts the number of schedulable resources participating in the resource bidding, it is specifically configured to: according to the state information of the schedulable resources participating in the resource bidding in the resource market, determine that it is necessary to increase the number of schedulable resources participating in the resource bidding. Bidding schedulable resources; among the schedulable resources that have never participated in the resource bidding, select the third schedulable resource that can participate in the resource bidding; and adjust the working status of the third schedulable resource to meet the requirements of being able to participate in the resource bidding to participate Resource bidding.

Further optionally, when the processor 52a determines that it is necessary to increase the schedulable resources that can participate in the resource bidding, it is specifically configured to: analyze the schedulable resources participating in the resource bidding according to the state information of the schedulable resources participating in the resource bidding in the resource market If the number of schedulable resources whose load is higher than the third load threshold is greater than the second number threshold, it is determined that it is necessary to increase the schedulable resources that can participate in resource bidding.

Furthermore, as shown in FIG. 5a, the resource control device further includes: a communication component 53a, a display 54a, a power supply component 55a, an audio component 56a and other components. Only some components are schematically shown in Fig. 5a, which does not mean that the resource control device only includes the components shown in Fig. 5a. In addition, the components in the dashed box in FIG. 5a are optional components, not mandatory components, and the specifics may depend on the product form of the resource control device. The resource control device in this embodiment can be implemented as a terminal device such as a desktop computer, a notebook computer, or a smart phone, or can be a server device such as a conventional server, a cloud server, or a server array. If the resource control device of this embodiment is implemented as a terminal device such as a desktop computer, a notebook computer, a smart phone, etc., it may include the components in the dashed box in FIG. 5a; if the resource control device of this embodiment is implemented as a conventional server, a cloud server, or a server Server devices such as arrays may not include the components in the dashed box in Figure 5a.

Correspondingly, an embodiment of the present application also provides a computer-readable storage medium storing a computer program, and when the computer program is executed, each step that can be executed by the resource control device in the foregoing method embodiment can be realized.

Fig. 5b is a schematic structural diagram of a resource application device provided by an exemplary embodiment of this application. As shown in FIG. 5b, the resource application device includes: a memory 51b, a processor 52b, and a communication component 53b.

The memory 51b is used to store computer programs, and can be configured to store other various data to support operations on the resource application device. Examples of these data include instructions for any application or method to operate on the resource application device, contact data, phone book data, messages, pictures, videos, etc.

The processor 52b is coupled with the memory 51b and is configured to execute a computer program for: determining at least one task that needs to be completed by the first application program and resource information that needs to be used by the at least one task; according to the resource information that needs to be used by the at least one task, Determine the actual bid value given for the resources required for at least one task application; generate a resource bid request based on the resource information required for at least one task and the actual bid value given for the resources required for the at least one task application; through the communication component 53b The resource bidding request is sent to the resource dispatcher for the resource dispatcher to bid for the first application and other applications to apply for schedulable resources; and to receive the result of the first application bidding for schedulable resources returned by the resource dispatcher.

In an optional embodiment, when the processor 52b determines the at least one task that needs to be completed by the first application program and the resource information that the at least one task needs to use, it is specifically configured to: use the operating law prediction model to learn the performance of the first application program. Operation law: According to the operation law of the first application program, at least one task to be completed by the first application program and resource information to be used by the at least one task are determined.

In an optional embodiment, when the processor 52b determines the actual bid value given for the resource required by the at least one task application, it is specifically configured to: according to the resource information required to be used by the at least one task, obtain information from the known schedulable resource information Determine the candidate schedulable resource and its reference bid value in the process; according to the total budget price of this resource application and the reference bid value of the candidate schedulable resource, determine the actual bid value given by the resources required for at least one task application.

In an optional embodiment, when the processor 52b generates a resource bidding request, it is specifically configured to: generate a resource bidding request for each task separately, and the resource bidding request for each task includes resource information that the task needs to use and a request for the task. The actual bid value given by the required resources; or, encapsulate the resource information required for at least one task and the actual bid value given for the required resources for at least one task application into the same resource bid request.

In an optional embodiment, after the processor 52b determines the actual competitive value given by the resource required for the at least one task application, it is also used to: run the controlled schedulable resource in the future period according to the feedback of the resource management and control party Status information to adjust the actual competitive value given for the resources required for at least one task application.

In an optional embodiment, when the processor 52b adjusts the actual competitive value given by the resource required for the at least one task application, it is specifically used to: according to the resource management and control party's feedback of the controlled schedulable resource in the future period of time Running status information, judging whether the number of resources required by at least one task is less than the third number threshold in the future period; if the judgment result is yes, increase the actual competition value given by the resources required by the at least one task application.

Furthermore, as shown in FIG. 5b, the resource application device further includes: a display 54b, a power supply component 55b, an audio component 56b and other components. Only some components are schematically shown in FIG. 5b, which does not mean that the resource application device only includes the components shown in FIG. 5b. In addition, the components in the dashed box in FIG. 5b are optional components, not mandatory components, and the specifics may depend on the product form of the resource application device. The resource application device in this embodiment may be implemented as a terminal device such as a desktop computer, a notebook computer, or a smart phone, or may be a server device such as a conventional server, a cloud server, or a server array. If the resource application device of this embodiment is implemented as a terminal device such as a desktop computer, notebook computer, smart phone, etc., it may include the components in the dashed box in Figure 5b; if the resource application device of this embodiment is implemented as a conventional server, cloud server or server Server devices such as arrays may not include the components in the dashed box in Figure 5b.

Correspondingly, the embodiment of the present application also provides a computer-readable storage medium storing a computer program, which can implement each step in the method embodiment shown in FIG. 4b when the computer program is executed.

FIG. 5c is a schematic structural diagram of another resource application device provided by an exemplary embodiment of this application. As shown in FIG. 5c, the resource application device includes: a memory 51c, a processor 52c, and a communication component 53c.

The memory 51c is used to store computer programs, and can be configured to store other various data to support operations on the resource application device. Examples of these data include instructions for any application or method to operate on the resource application device, contact data, phone book data, messages, pictures, videos, etc.

The processor 52c, coupled with the memory 51c, is used to execute computer programs for: determining the number of resources that the first application needs to apply for this time and the total budget of this resource application; according to the amount of resources that need to be applied for this time and The total budget of this resource application generates a resource bid request; the resource bid request is sent to the resource scheduler through the communication component 53c, so that the resource scheduler can bid for the first application and other applications to apply for schedulable resources and return; And receiving the result of the first application bidding for schedulable resources returned by the resource scheduler.

In an optional embodiment, when the processor 52c determines the amount of resources that the first application program needs to apply for this time, it is specifically configured to: determine at least one task that the first application program needs to complete according to the running law of the first application program And resource information that needs to be used by at least one task; according to the resource information that needs to be used by at least one task, the number of resources that the first application needs to apply for this time is determined.

Further, as shown in FIG. 5c, the resource application device further includes: a display 54c, a power supply component 55c, an audio component 56c and other components. Only some components are schematically shown in FIG. 5c, which does not mean that the resource application device only includes the components shown in FIG. 5c. In addition, the components in the dashed box in FIG. 5c are optional components, not mandatory components, and the specifics may depend on the product form of the resource application device. The resource application device in this embodiment can be implemented as a terminal device such as a desktop computer, a notebook computer, or a smart phone, or it can be a server device such as a conventional server, a cloud server, or a server array. If the resource application device of this embodiment is implemented as a terminal device such as a desktop computer, a notebook computer, a smart phone, etc., it may include the components in the dashed box in Figure 5c; if the resource application device of this embodiment is implemented as a conventional server, a cloud server or a server Server devices such as arrays may not include the components in the dashed box in Figure 5c.

Correspondingly, an embodiment of the present application also provides a computer-readable storage medium storing a computer program, which can implement each step in the method embodiment shown in FIG. 4c when the computer program is executed.

Fig. 5d is a schematic structural diagram of a resource scheduling device provided by an exemplary embodiment of this application. As shown in FIG. 5d, the resource scheduling device includes: a memory 51d, a processor 52d, and a communication component 53d.

The memory 51d is used to store computer programs and can be configured to store various other data to support operations on the resource scheduling device. Examples of these data include instructions for any application or method operating on the resource scheduling device, contact data, phone book data, messages, pictures, videos, etc.

The processor 52d, coupled with the memory 51d, is configured to execute a computer program for: receiving a resource bid request through the communication component 53d. The resource bid request includes: the number of resources that the first application needs to apply for this time and the amount of this resource application Total budget price; Determine the actual competitive value of at least one resource to be applied for according to the number of resources to be applied for this time and the total budget price of this resource application; According to the actual competitive value of at least one resource to be applied for, it is the first application and Other applications bid for at least one resource to be applied for, and the resource application result of the first application is returned through the communication component.

In an optional embodiment, when determining at least one resource to be applied for and its actual bid value, the processor 52d is specifically configured to: determine at least one of the schedulable resources participating in resource bidding according to the number of resources to be applied for this time A resource to be applied for; the actual competitive value of at least one resource to be applied for is determined based on the total budget price of this resource application and the reference bid value of at least one resource to be applied for.

In an optional embodiment, when the processor 52d determines the actual bid value of at least one resource to be applied for, it is specifically configured to: determine at least one of the schedulable resources participating in the resource bidding according to the number of resources to be applied for this time Resources to be applied for; based on the number of resources to be applied for this time and the total budget of this resource application, the average bid value is calculated as the actual bid value of at least one resource to be applied for.

Further, as shown in FIG. 5d, the resource scheduling device further includes: a display 54d, a power supply component 55d, an audio component 56d and other components. Only some components are schematically shown in FIG. 5d, which does not mean that the resource scheduling device only includes the components shown in FIG. 5d. In addition, the components in the dashed box in FIG. 5d are optional components, not mandatory components, and the specifics may depend on the product form of the resource scheduling device. The resource scheduling device of this embodiment can be implemented as a terminal device such as a desktop computer, a notebook computer, or a smart phone, or can be a server device such as a conventional server, a cloud server, or a server array. If the resource scheduling device of this embodiment is implemented as a terminal device such as a desktop computer, notebook computer, smart phone, etc., it may include the components in the dashed box in Figure 5d; if the resource scheduling device of this embodiment is implemented as a conventional server, a cloud server or a server Server devices such as arrays may not include the components in the dashed box in Figure 5d.

Correspondingly, an embodiment of the present application also provides a computer-readable storage medium storing a computer program, and when the computer program is executed, each step that can be executed by the resource scheduling device in the foregoing method embodiment can be implemented.

The above-mentioned memory in Figures 5a-5d can be implemented by any type of volatile or non-volatile storage device or their combination, such as static random access memory (SRAM), electrically erasable programmable read-only memory ( EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

The communication components in Figures 5a-5d described above are configured to facilitate wired or wireless communication between the device where the communication component is located and other devices. The device where the communication component is located can access wireless networks based on communication standards, such as WiFi, 2G, 3G, 4G/LTE, 5G and other mobile communication networks, or a combination of them. In an exemplary embodiment, the communication component receives a broadcast signal or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component may further include a near field communication (NFD) module, radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra wideband (UWB) technology, Bluetooth (BT) technology Wait.

The above-mentioned display in Figs. 5a-5d includes a screen, and the screen may include a liquid crystal display (LDD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, sliding, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure related to the touch or slide operation.

The power supply components in Figures 5a to 5d above provide power for various components of the equipment where the power supply component is located. The power supply component may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device where the power supply component is located.

The audio components in Figs. 5a to 5d may be configured to output and/or input audio signals. For example, the audio component includes a microphone (MID). When the device where the audio component is located is in an operating mode, such as call mode, recording mode, and voice recognition mode, the microphone is configured to receive external audio signals. The received audio signal can be further stored in a memory or sent via a communication component. In some embodiments, the audio component further includes a speaker for outputting audio signals.

Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

The present invention is described with reference to flowcharts and/or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and/or block in the flowchart and/or block diagram, and the combination of processes and/or blocks in the flowchart and/or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing equipment to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing equipment are generated It is a device that realizes the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.

These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing equipment to work in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction device. The device implements the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.

These computer program instructions can also be loaded on a computer or other programmable data processing equipment, so that a series of operation steps are executed on the computer or other programmable equipment to produce computer-implemented processing, so as to execute on the computer or other programmable equipment. The instructions provide steps for implementing the functions specified in one process or multiple processes in the flowchart and/or one block or multiple blocks in the block diagram.

In a typical configuration, the computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include non-permanent memory in a computer readable medium, random access memory (RAM) and/or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). Memory is an example of computer readable media.

Computer-readable media include permanent and non-permanent, removable and non-removable media, and information storage can be realized by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, CD-ROM, digital versatile disc (DVD) or other optical storage, Magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media can be used to store information that can be accessed by computing devices. According to the definition in this article, computer-readable media does not include transitory media, such as modulated data signals and carrier waves.

It should also be noted that the terms "include", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or equipment including a series of elements not only includes those elements, but also includes Other elements that are not explicitly listed, or they also include elements inherent to such processes, methods, commodities, or equipment. If there are no more restrictions, the element defined by the sentence "including a..." does not exclude the existence of other identical elements in the process, method, commodity, or equipment that includes the element.

The foregoing descriptions are only examples of the present application, and are not used to limit the present application. For those skilled in the art, this application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this application shall be included in the scope of the claims of this application.

Claims

A system, characterized by comprising: an application layer, an intermediate layer, and an IT resource layer; the intermediate layer is between the application layer and the IT resource layer, and is used to adopt a resource bidding and regulation mechanism to combine all The resource requirements of the application layer perform resource scheduling on the IT resource layer.
The system according to claim 1, wherein the IT resource layer includes at least one schedulable IT resource; the intermediate layer includes: a resource scheduling node and a resource control node;

The resource control node is used to adopt a resource control mechanism to control IT resources that can participate in resource bidding in the IT resource layer;

The resource scheduling node is configured to use a resource bidding mechanism to schedule IT resources in the IT resource layer that can participate in resource bidding in combination with the resource requirements of the application layer.
The system according to claim 2, wherein the middle layer further comprises: a resource manager;

The resource manager is used to control the power consumption of the IT resources in the IT resource layer, and feed back the running status of the related IT resources to the resource control node.
The system according to claim 3, wherein the resource manager is specifically configured to: use the first machine learning model to learn the relationship between the load and power consumption of each IT resource, and according to the learned relationship between the load and the power consumption , Perform power consumption control on each IT resource.
The system according to claim 3, wherein the resource control node is specifically used to combine the running status of related IT resources fed back by the resource manager, and use a second machine learning model to learn how IT resources participate in resource bidding. The law of change: A resource control mechanism is used, combined with the learned law of change, to control the IT resources in the IT resource layer that can participate in resource bidding.
The system according to claim 3, wherein the intermediate layer further comprises: a proxy node for initiating a resource bidding request to the resource scheduling node for an application in the application layer to apply for resources.
The system according to claim 6, wherein the proxy node is specifically configured to:

Use the third machine learning model to learn the operating law of the application program corresponding to the agent node;

Determine at least one task that needs to be completed by the application program corresponding to the agent node and resource information that the at least one task needs to use according to the operating law of the application program corresponding to the agent node;

According to the resource information that the at least one task needs to use, a resource bidding request of the application program corresponding to the agent node is generated.
A resource scheduling method, characterized in that it comprises:

Obtain the resource requirements of upper-level applications;

The resource bidding and regulation mechanism is adopted, combined with the resource requirements of the upper-layer applications, to perform resource scheduling on the lower-layer IT resources.
The method according to claim 8, characterized in that, using a resource bidding and regulation mechanism and combining the resource requirements of the upper-layer application to perform resource scheduling on the lower-layer IT resources comprises:

Adopt a resource control mechanism to control IT resources that can participate in resource bidding among lower-level IT resources;

A resource bidding mechanism is used to schedule IT resources that can participate in resource bidding among the lower-level IT resources in combination with the resource requirements of the upper-layer application.
The method according to claim 9, characterized in that, adopting a resource control mechanism to regulate and control IT resources that can participate in resource bidding among lower-level IT resources, comprising:

A resource control mechanism is adopted, combined with the operating status of related IT resources in the lower-level IT resources, to regulate and control the IT resources in the lower-level IT resources that can participate in resource bidding.
A resource control method, which is characterized in that it comprises:

Monitor the status information of schedulable resources participating in resource bidding in the resource market;

According to the state information of the schedulable resources participating in the resource bidding in the resource market, the schedulable resources participating in the resource bidding are regulated and controlled.
The method according to claim 11, characterized in that, according to the state information of the schedulable resources participating in the resource bidding in the resource market, the regulation of the schedulable resources participating in the resource bidding includes at least one of the following regulation operating:

Adjust the actual bid value of the schedulable resource participating in the resource bidding in the resource market according to the state information of the schedulable resource participating in the resource bidding in the resource market;

According to the state information of the schedulable resources participating in the resource bidding in the resource market, the quantity of the schedulable resources participating in the resource bidding is adjusted.
The method according to claim 12, wherein the actual bid value of the schedulable resource participating in the resource bidding in the resource market is adjusted according to the state information of the schedulable resource participating in the resource bidding in the resource market, include:

According to the state information of the schedulable resources participating in the resource bidding in the resource market, determine the first schedulable resource whose bid value needs to be adjusted;

A first resource bidding request is sent to the resource scheduler, where the first resource bidding request includes a first actual bid value for adjusting the actual bid value of the first schedulable resource in the resource market.
The method according to claim 13, wherein, according to the state information of the schedulable resources participating in the resource bidding in the resource market, determining the first schedulable resource that needs to adjust the bid value comprises:

Calculating the benchmark bid value of the schedulable resource participating in the resource bidding according to the resource parameter that characterizes the operating state of the schedulable resource participating in the resource bidding in the resource market;

Obtaining the transaction bid value of the schedulable resource participating in the resource bidding from the bidding parameter that characterizes the bidding status of the schedulable resource participating in the resource bidding in the resource market;

According to the benchmark bid value and the transaction bid value of the schedulable resources participating in the resource bidding, the first schedulable resource whose bid value needs to be adjusted is determined.
The method according to claim 14, wherein the benchmark bid value of the schedulable resource participating in the resource bidding is calculated according to the resource parameter characterizing the operating state of the schedulable resource participating in the resource bidding in the resource market, include:

The resource parameter that characterizes the operating state of the schedulable resource participating in the resource bidding in the resource market is input into a cost prediction model, and the cost prediction model is used to calculate the benchmark bid value of the schedulable resource participating in the resource bidding.
The method according to claim 14 or 15, characterized in that, according to the benchmark bid value and the transaction bid value of the schedulable resources participating in the resource bidding, determining the first schedulable resource whose bid value needs to be adjusted includes:

From the schedulable resources participating in the resource bidding, a schedulable resource whose transaction bid value is less than a reference bid value is obtained as the first schedulable resource; wherein, the first actual bid value is greater than the reference bid value.
The method according to claim 12, wherein, according to the state information of the schedulable resources participating in the resource bidding in the resource market, adjusting the quantity of the schedulable resources participating in the resource bidding comprises:

According to the state information of the schedulable resources participating in the resource bidding in the resource market, determine the second schedulable resource that needs to withdraw from the resource bidding;

Sending a second resource bidding request to the resource scheduler, where the second resource bidding request includes a second actual bid value for applying for the second schedulable resource;

After the second schedulable resource is applied for, the working state of the second schedulable resource is adjusted to not meet the requirement of being able to participate in the resource bidding to withdraw from the resource bidding.
The method according to claim 17, wherein, according to the state information of the schedulable resources participating in the resource bidding in the resource market, determining the second schedulable resource that needs to withdraw from the resource bidding comprises:

Analyzing the load of the schedulable resource participating in the resource bidding in the resource market according to the state information of the schedulable resource participating in the resource bidding;

From the schedulable resources participating in the resource bidding, obtain a schedulable resource whose load is lower than a first load threshold as the second schedulable resource.
The method according to claim 18, wherein before acquiring a schedulable resource whose load is lower than a first load threshold as the second schedulable resource, the method further comprises:

Judging whether the number of schedulable resources whose load is lower than the second load threshold is greater than the first number threshold;

If the judgment result is yes, obtain a schedulable resource whose load is lower than the first load threshold from the schedulable resources participating in the resource bidding as the second schedulable resource;

Wherein, the second load threshold is greater than the first load threshold.
The method according to any one of claims 17-19, wherein adjusting the working state of the second schedulable resource to a requirement of being able to participate in resource bidding to withdraw from the resource bidding comprises:

Close the second schedulable resource; or

Controlling the second schedulable resource to enter a low power consumption mode not higher than a specified level; or

Adjusting the working frequency of the second schedulable resource to not higher than a specified frequency;

Among them, the designated level or designated frequency corresponds to not meeting the requirement of being able to participate in resource bidding.
The method according to claim 12, wherein, according to the state information of the schedulable resources participating in the resource bidding in the resource market, adjusting the quantity of the schedulable resources participating in the resource bidding comprises:

According to the state information of the schedulable resources participating in the resource bidding in the resource market, it is determined that it is necessary to increase the schedulable resources that can participate in the resource bidding;

Among the schedulable resources that have never participated in the resource bidding, select the third schedulable resource that can participate in the resource bidding; and

The working state of the third schedulable resource is adjusted to meet the requirement of being able to participate in the resource bidding to participate in the resource bidding.
The method according to claim 21, characterized in that, according to the state information of the schedulable resources participating in the resource bidding in the resource market, determining that it is necessary to increase the schedulable resources participating in the resource bidding comprises:

Analyzing the load of the schedulable resource participating in the resource bidding in the resource market according to the state information of the schedulable resource participating in the resource bidding;

If the number of schedulable resources whose load is higher than the third load threshold is greater than the second number threshold, it is determined that it is necessary to increase the schedulable resources that can participate in resource bidding.
A resource application method, characterized in that it includes:

Determine at least one task that needs to be completed by the first application program and resource information that the at least one task needs to use;

Determine the actual competitive value given for the resource required by the at least one task application according to the resource information that the at least one task needs to use;

Generating a resource bid request according to the resource information that the at least one task needs to use and the actual bid value given for the resource required for the at least one task application;

Sending the resource bid request to a resource scheduler, so that the resource scheduler can bid for the first application and other applications to apply for schedulable resources;

Receiving a result of the first application bidding for a schedulable resource returned by the resource scheduler.
The method according to claim 23, wherein determining at least one task to be completed by the first application program and resource information to be used by the at least one task comprises:

Use the operation rule prediction model to learn the operation rule of the first application program;

According to the running law of the first application program, determine at least one task to be completed by the first application program and resource information to be used by the at least one task.
The method according to claim 23, wherein, according to the resource information that the at least one task needs to use, determining the actual competitive value given for the resource required by the at least one task application comprises:

Determine candidate schedulable resources and their reference competitive value from the known schedulable resource information according to the resource information that the at least one task needs to use;

According to the total budget price of this resource application and the reference bid value of the candidate schedulable resource, the actual bid value given for the resource required by the at least one task application is determined.
The method according to claim 23, wherein generating the resource bid request according to the resource information that the at least one task needs to use and the actual bid value given for the resource required by the at least one task application includes :

A resource bidding request is generated for each task separately, and the resource bidding request for each task includes the resource information that the task needs to use and the actual bid value given for the resources required by the task request; or

The resource information that the at least one task needs to use and the actual bid value given for the resource required by the at least one task application are encapsulated into the same resource bidding request.
The method according to claim 26, wherein after determining the actual competitive value given for the resource required for the at least one task application, the method further comprises:

According to the operating state information of the managed and schedulable resource fed back by the resource management and control party in the future period of time, the actual competitive value given by the resource required for the at least one task application is adjusted.
The method according to claim 27, characterized in that, according to the running status information of the managed and schedulable resource fed back by the resource management and control party in the future period, the actual competitive value given to the resource required for the at least one task application Make adjustments, including:

Judging whether the number of resources required by the at least one task in the future period is less than the third number threshold in the future period according to the running state information of the managed and schedulable resource fed back by the resource management and control party in the future period;

If the judgment result is yes, increase the actual competitive value given for the resources required for the at least one task application.
A resource application method, characterized in that it includes:

Determine the number of resources that the first application needs to apply for this time and the total budget for this resource application;

Generate a resource bid request based on the number of resources to be applied for this time and the total budget price of this resource application;

Sending the resource bid request to a resource scheduler, so that the resource scheduler can bid for the first application and other applications to apply for schedulable resources;

Receiving a result of the first application bidding for a schedulable resource returned by the resource scheduler.
The method according to claim 29, wherein determining the number of resources that the first application program needs to apply for this time comprises:

Determine at least one task that needs to be completed by the first application program and resource information that the at least one task needs to use according to the running law of the first application program;

According to the resource information that the at least one task needs to use, the number of resources that the first application needs to apply for this time is determined.
A resource scheduling method, characterized in that it comprises:

Receiving a resource bid request, where the resource bid request includes: the number of resources that the first application needs to apply for this time and the total budget price of this resource application;

According to the number of resources to be applied for this time and the total budget price of this resource application, determine at least one resource to be applied for and its actual bid value among the schedulable resources participating in the resource bidding;

According to the actual bid value of the at least one resource to be applied for, bid for the at least one resource to be applied for for the first application and other applications, and return the resource application result of the first application.
The method according to claim 31, characterized in that, based on the number of resources to be applied for this time and the total budget price of this resource application, among the schedulable resources participating in the resource bidding, at least one resource to be applied for and its actual value are determined. Competitive value, including:

According to the number of resources to be applied for this time, at least one resource to be applied for is determined among the schedulable resources participating in the resource bidding;

Determine the actual bid value of the at least one resource to be applied for according to the total budget price of the resource application and the reference bid value of the at least one resource to be applied for.
The method according to claim 32, wherein the reference bid value of the at least one resource to be applied for comprises: the benchmark bid value of the at least one resource to be applied for, the average actual bid value in the resource market, or the reference bid value of the at least one resource to be applied for. The largest actual competitive value in the market.
The method according to claim 31, characterized in that determining the actual competitive value of at least one resource to be applied for based on the number of resources to be applied for this time and the total budgeted price of this resource application, including:

According to the number of resources to be applied for this time, at least one resource to be applied for is determined among the schedulable resources participating in the resource bidding;

According to the number of resources to be applied for this time and the total budgeted price of this resource application, the average bid value is calculated as the actual bid value of the at least one resource to be applied for.
A resource control device, which is characterized by comprising: a memory and a processor;

The memory is used to store a computer program;

The processor, coupled with the memory, is configured to execute the computer program for:

Monitor the status information of schedulable resources participating in resource bidding in the resource market;

According to the state information of the schedulable resources participating in the resource bidding in the resource market, the schedulable resources participating in the resource bidding are regulated and controlled.
A resource application device, characterized by comprising: a memory, a processor, and a communication component;

The memory is used to store a computer program;

The processor, coupled with the memory, is configured to execute the computer program for:

Determine at least one task that needs to be completed by the first application program and resource information that the at least one task needs to use;

Determine the actual competitive value given for the resource required by the at least one task application according to the resource information that the at least one task needs to use;

Generating a resource bid request according to the resource information that the at least one task needs to use and the actual bid value given for the resource required for the at least one task application;

Send the resource bid request to the resource scheduler through the communication component, so that the resource scheduler can bid for the first application with other applications for schedulable resources; and receive the return from the resource scheduler The result of the first application bidding for schedulable resources.
A resource application device, characterized by comprising: a memory, a processor, and a communication component;

The memory is used to store a computer program;

The processor, coupled with the memory, is configured to execute the computer program for:

Determine the number of resources that the first application needs to apply for this time and the total budget for this resource application;

Generate a resource bid request based on the number of resources to be applied for this time and the total budget price of this resource application;

Send the resource bid request to the resource scheduler through the communication component, so that the resource scheduler can bid for the first application and other applications for schedulable resources and return; and receive the resource scheduler The returned result of the first application bidding for schedulable resources.
A resource scheduling device, which is characterized by comprising: a memory, a processor, and a communication component;

The memory is used to store a computer program;

The processor, coupled with the memory, is configured to execute the computer program for:

Receive a resource bid request through the communication component, where the resource bid request includes: the number of resources that the first application needs to apply for this time and the total budget price of this resource application;

Determine the actual competitive value of at least one resource to be applied for according to the number of resources to be applied for this time and the total budget price of this resource application;

According to the actual bid value of the at least one resource to be applied for, bid for the at least one resource to be applied for for the first application and other applications, and return the resource application of the first application through the communication component result.
A resource management system, characterized by comprising: a resource scheduling party, a resource demanding party, and a resource regulating party;

The resource demander is configured to generate a resource bidding request for the first application; send the resource bidding request to the resource scheduler to participate in the resource bidding;

The resource scheduler is configured to apply for schedulable resources for the first application program and other application programs in the schedulable resources participating in the resource bidding according to the resource scheduling request;

The resource regulator is used to monitor the status information of the schedulable resources participating in the resource bidding in the resource market; according to the status information of the schedulable resources participating in the resource bidding in the resource market, check the status information of the schedulable resources participating in the resource bidding Scheduling resources for regulation.
A computer room system, characterized by comprising: at least one schedulable resource, at least one infrastructure, resource application equipment, resource scheduling equipment, and resource control equipment; the at least one infrastructure provides the basis for the at least one schedulable resource service;

The resource demand device is configured to generate a resource bidding request for a first application program; send the resource bidding request to the resource scheduling device to participate in the resource bidding;

The resource scheduling device is configured to, according to the resource scheduling request, apply for a schedulable resource for the first application program and other application programs among the schedulable resources participating in the resource bidding;

The resource control device is used to monitor the status information of the schedulable resources participating in the resource bidding in the resource market; according to the status information of the schedulable resources participating in the resource bidding in the resource market, check the status information of the schedulable resources participating in the resource bidding Scheduling resources for regulation.
A data center system is characterized by comprising: at least one computer room, resource application equipment, resource scheduling equipment, and resource control equipment; each computer room includes at least one schedulable resource and at least one infrastructure, and the at least one infrastructure is The at least one schedulable resource provides basic services;

The resource demand device is configured to generate a resource bidding request for a first application program; send the resource bidding request to the resource scheduling device to participate in the resource bidding;

The resource scheduling device is configured to, according to the resource scheduling request, apply for a schedulable resource for the first application program and other application programs among the schedulable resources participating in the resource bidding;

The resource control device is used to monitor the status information of the schedulable resources participating in the resource bidding in the resource market; according to the status information of the schedulable resources participating in the resource bidding in the resource market, check the status information of the schedulable resources participating in the resource bidding Scheduling resources for regulation.
A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor is caused to implement the steps in the method of any one of claims 1-34.